Which books? Where? ...

Enhancing an academic library
renovation project with creative
open stack print collections services
Lorrie McAllister, Shari Laster, Tammy Dang, Emily Pattni

June 2024

This work made possible by the Andrew W. Mellon Foundation.

Enhancing an academic library renovation project

1 of 47

Abstract
This paper describes how Arizona State University Library used creativity and novel
approaches to collections design and implementation processes to select open
stack print books for a newly renovated academic research library. Using results
from a workshop focused on rethinking the future of print within educational learning
and research environments, the Collections Services and Analysis unit within Arizona
State University Library performed a series of experiments to better understand
the purpose and use of print collections within 21st century library design. The
authors describe the creative processes used in collections design and three types
of selection approaches that invited engagement with open stacks. These three
types were: small browsing collections co-curated with community members, a
medium-sized print collection selected for student engagement, and a large research
collection selected using a novel data analysis of four factors affecting the likelihood
of potential use. Using more than one million volumes as the basis for selection,
approximately 185,000 volumes were installed in the renovated library through a
complex implementation across four library locations. The authors discuss the key
role that creativity played in the approaches, methods, and results of these efforts
and offer recommendations for collection management teams seeking to maximize
their pursuit of community engagement with print collections within contemporary
academic library spaces.

Keywords
collection selection, creativity in libraries, print book collections, open stacks,
data analysis, collection development, library spaces, library shelving, library
planning, open stacks, library renovation, collection design, design process,
collections engagement

Enhancing an academic library renovation project

2 of 47

Table of Contents

Introduction.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 4
University Library context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  5
Hayden Library reinvention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  5
Reinventing Hayden Library collections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  7
Selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Materials and methods .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 11
Overall approach to shelving. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  12
Figure 1: S
 ample elevation drawing of perimeter shelving. . . . . . . . . . .  12
Small featured collections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  13
Student-focused collections - Sun Devil Reads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  14
Designing test-fits based on shelving capacity .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 15
Table 1: Shelving style configurations and estimated volume counts for
Sun Devil Reads category sections.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  17
Figure 2: S
 ample category layout for Sun Devil Reading shelving.. . .  18
Large-scale open stacks for research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  19
Assessing ASU Library’s collections.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 19
Flowing, scaling and selecting.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 20
Figure 3: E
 xamples of potential flows for putting LC call numbers
in order in the fourth floor stacks. Note that these diagrams are in
reverse to the final layout.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  21
Figure 4: F
 inal flow for putting LC call numbers in order in the fourth
floor stacks.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Table 2: Example of an A/B testing choice.. . . . . . . . . . . . . . . . . . . . . . . . 24
Table 3: E
 xample demonstrating how criteria are applied to order and
rank books in a set.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Sample criteria for a single subclass.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 25
Table 4: S
 ample output of stepwise criteria. These criteria were
adjusted for each LC subclass to produce the approximate count of
items corresponding to the estimated capacity of the desired location
within the overall layout.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Selection in practice. .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 27
“Cascading Windows of Sets”.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 28

Enhancing an academic library renovation project

3 of 47

Results .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 29
Preparing books to move. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Figure 5: S
 ample of start and stop flags. . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Preparing destinations.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 30
Figure 6: E
 xample of floor map indicating the destination locations for
LC subclass sections.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Moving phases and process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Discussion.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 34
Selection and development of collections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Displaying collections in libraries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Analyzing and using data for collections design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Renovating operations and workflows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
The New Open Stacks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Acknowledgements .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 38
References .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 39
Appendix.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 41

Enhancing an academic library renovation project

4 of 47

Introduction
In 2017, Arizona State University (ASU) embarked on an approximately two-year, $90
million reinvention project to transform Hayden Library, a flagship library focused
on humanities and social sciences, to support 21st century learning and discovery.
Senior ASU Library leadership reconceptualized the library’s spaces and services
through a massive building renovation, which opened the door to a fresh approach
to the design of library collections within this building.1 This paper describes how the
ASU Library Collections Services and Analysis team used creativity to address the
design challenges inherent in planning and implementing physical collections for a
reconceptualized, contemporary, 252,600-square-foot academic library, in time for
its reopening in early 2020.
The scholarly literature of the past two decades elucidates a sea change in library
design that significantly reduces or eliminates print collections in libraries in favor
of repurposing space to accommodate the changing social needs of teachers,
learners, and researchers (Bennett, 2003; Somerville & Brown-Sica, 2011; Sandy,
Krishnamurthy & Scalfani, 2014; Yi, 2015). While choosing to follow this design
trend, senior leaders, including University Librarian Jim O’Donnell, also recognized
the value in retaining print collections within academic libraries (O’Donnell, 2017).
Keeping print materials within a user-centered dynamic space called for the
exploration of new strategies for using print materials as a method to engage,
educate, and inspire users in a contemporary educational context. Print books are
no longer “pleasant furnishings and backdrops for quiet study,” but may function
as active objects of engagement for a new generation of students and researchers
(McAllister & Laster, 2018). Grant funding supported a project team tasked with the
design and selection of inclusive and engaging print collections for library visitors
in the newly renovated facility. As part of this “future of print’’ initiative, library staff
were empowered to use their creativity to develop and implement workflows to
select, acquire, stage, relocate, and manage open stack collections that aligned
with the reinvention project. The project team subsequently developed three
approaches to selection of materials. Developing a clear vision and narratives about
the purpose of collections on each floor of the library and then employing these
approaches facilitated designing collections for a massive, empty library. This paper
describes the creative design process and approaches to open stack book selection
methods, including a novel data analysis process for a large research collection.
The authors communicate the results of these processes, along with an overview
of the complex three-phase implementation, resulting in the relocation of around
830,000 circulating volumes among four library locations over a twelve-month
1

Design can be used in a myriad of ways. Here, we use the word “design” to mean the “activity
of imagining the future, setting goals, and finding ways to achieve those goals,” as described by
Jackson (2009).

Enhancing an academic library renovation project

5 of 47

period. The discussion describes the role of creativity in library collections work and
the lessons learned regarding the three approaches to selection, implementation,
and initial reception of the collections. The authors also provide recommendations
for collections leadership in using print as an engagement tool within contemporary
library spaces.
University Library context
At the point the project planning began in 2017, Arizona State University served
nearly 73,000 students in its full-immersion programs on four campuses in the
Phoenix metropolitan area, and an additional 30,500 students in its online programs;
for Fall 2023 these numbers are nearly 80,000 and over 66,000, respectively.2 The
university’s decades of growth is the result of President Michael Crow’s expansive
vision to significantly broaden access to high quality public education while also
contributing to meaningful societal impact (Crow & Dabars, 2015; 2020). ASU
Library operates within this conceptual space, seeking to align its services and
practices with the New American University concept by expanding its online services
while intentionally designing flexible campus library spaces that support a variety
of academic and intellectual engagements, as well as quiet study and reflection
(McAllister & Laster, 2018a). ASU Library serves all students in the 800+ academic
undergraduate and graduate programs and majors available at the university, along
with the teaching and research needs of the faculty. The Library now hosts eight
library locations, four of which are on the Tempe campus, including the largest
library on this campus, Hayden Library. Prior to the renovation project, Hayden
Library was home to 1.1 million volumes, including Library of Congress (LC) classes
for all disciplines except the arts, music, and science and technology. The secondlargest library, Noble Library, was constructed in 1983 as the campus science and
engineering library. The other two locations on the Tempe campus, the Design and
the Arts Library and the Music Library, offer spaces and services for engagement
with these disciplinary communities. In total, ASU Library locations have more than
two million visitors to our buildings annually and house millions of print books, in
addition to extensive online holdings. Library collections play an important role within
the expansive concept of the New American University, which informed the design
approach to collections within the Hayden Library reinvention project.
Hayden Library reinvention
Hayden Library, named for Charles Trumbull Hayden, a colonial settler, businessman,
and judge who helped found the city of Tempe and Arizona State University, was
originally built in 1966 by the architectural firm of Frederick Weaver and Richard
Drover. The building consisted of a four-story “tower” and housed 600,000 printed
2

For enrollment trends, see: https://www.asu.edu/facts/#/facts/enrollment/campus-major.

Enhancing an academic library renovation project

6 of 47

volumes that were transferred from the old Matthews Library to serve the needs
of humanities and social sciences learners and scholars at the growing Tempe
campus (Arizona Board of Regents, 2024b). The library was subsequently expanded
in 1989 to meet the needs of the growing student population at ASU. The scope
of the 2018-2020 Hayden Library renovation included only the 1966 portion of the
building. The 1989-era lower concourse and lower level remained open throughout
the renovation project, with staff workspaces, University classrooms, and a small
collections space. The lowest floor of the 1966 building was the last section of
Hayden to close for renovation in early 2018, and the first to reopen post-renovation,
with University classrooms scheduled for use in Fall 2019 and subsequent floors
scheduled to open early in 2020. The closing and reopening timelines for these
building sections necessitated the development of a complex collections relocation
plan and staggered collection development planning that matched the building’s
renovation timeline.
The University sought to update Hayden Library and reinvent it as “a hub of
21st century discovery, innovation and learning” that would help “materialize the
vision of ASU as the New American University and the leading center for global
interdisciplinary scholarship and discovery” (Arizona Board of Regents, 2024). This
reinvention project offered an opportunity to rethink how physical collections may be
used to meet the contemporary needs of researchers, scholars, and learners who
largely work within digital scholarship contexts.
Starting in the summer of 2017, well over one million books were dispersed among
other library locations to clear the 1966 building for the renovation project. The
1989 building addition remained open as a service point and to accommodate a
limited number of readers and classes. The collections materials that were cleared
from Hayden Library included federal, state, and local government documents,
microforms, and specialized holdings, including rare books and archival materials,
which were sent to other libraries or off-site storage facilities, depending on
collection’s preservation and access needs, content focus, LC subclass, and/or loan
frequency. For example, books that fell into the higher-circulating LC subclasses,
such as Language and Literature, were sent to Noble Library, just a short walk from
Hayden, while the bulk of the STEM collections that had been housed at Noble
Library were sent to temporary storage on the Polytechnic campus for processing
into the library’s high-density storage facility. Books that fell into less-circulated
LC subclasses were also processed into high-density storage. The extensive
microforms collection was relocated to a temporary space on the Polytechnic
campus to facilitate microform retrieval and digitization services. State and local
government documents that had records in the library catalog were processed into
high-density storage, while federal documents were relocated to the West Valley
campus in Glendale and housed at Fletcher Library to facilitate retrieval. Library staff
also inventoried and moved the vast majority of specialized collections holdings from

Enhancing an academic library renovation project

7 of 47

Hayden Library to a new, permanent storage location. In all, more than one million
books and more than 2,300 linear feet of special collections were removed from
Hayden Library. It was a period of massive change for ASU Library, Tempe campus,
and the local ASU community.
Reinventing Hayden Library collections
University Librarian Jim O’Donnell secured a planning grant from The Andrew
W. Mellon Foundation to support intra-institutional dialogue about “the future
of print” in academic libraries. Lorrie McAllister, Associate University Librarian,
served as project manager for this grant, developing the collaboratively authored
white paper “The Future of the Academic Library Print Collection: A Space for
Engagement,” released in October 2017 (ASU Library, 2021; ASU Library, 2017). With
the knowledge gained from these conversations, McAllister drafted a three-year,
$381,000 implementation grant titled “The Future of the Arizona State University
Library Print Collection: A Collaborative and Data-driven Approach to Stack Design
and Curation” in coordination with O’Donnell, which was also funded by The Andrew
W. Mellon Foundation. The subsequent project work necessitated experimentation
with collections as an engagement tool, conceptual work regarding how to conceive
of physical collections within a large 21st century academic research library, and
creative design and process work to inform which collections would reenter the
newly renovated library.
Following the white paper, and with the implementation grant funding, McAllister
began assembling a team to successfully carry out the collections portion of the
Hayden Library reinvention project. In January 2018, Shari Laster joined ASU Library
as the Head of Open Stack Collections, a new unit that united operations and
strategic program implementation for all of the library’s open stack collections in
print and digital formats. Laster then served as the project coordinator for the Mellon
Foundation-funded “future of print” grant under McAllister’s leadership. Tammy
Dang, Data Analysis Specialist, and Emily Pattni, Curation Specialist, were hired in
the summer of 2018. With this team, and with the collaboration of many other library
staff members across the organization, we began our work.
First, we experimented with reconceiving open stack collections as an engagement
tool through small-scale browsable collections. The team developed eight
collections, pairing a library user group with a potential space, a subject matter or
concept, and a method to select materials for the collection. For example, Pattni led
the creation of a collection of health resources in collaboration with faculty members
in the Health Sciences for the Downtown Library, using a shared list-making
approach to selection. The team then spent almost a year in collaborative processes
to select books and materials for these small browsing collections and learn from
the methods used, lessons learned, and impacts on users, while understanding the

Enhancing an academic library renovation project

8 of 47

resources needed to continue this “high touch” collections service past the grant
period.3 This experimentation resulted in our later development of the Sun Devil
Reads collection for Hayden Library, as mentioned later in this paper.
With our project team in place and some community engagement underway,
we embarked on the conceptual work of designing Hayden Library collections
for the reinvention project. The architecture firm for the reinvention, Ayers Saint
Gross, and the project manager for ASU Library, Associate University Librarian
Tomalee Doan, provided McAllister with initial floor layouts, shelving diagrams, and
approximate linear foot counts for each of the five floors. Due to deep experience
with contemporary learning space design, Doan led space planning for the
library, specifying flexible, reconfigurable classrooms and collaboration spaces, a
makerspace, a geospatial hub, a data science instruction space, and several other
updates to the library that brought the space use into an updated research context.
McAllister, meanwhile, conceived of the collections design, working within the project
specifications, overall vision for the reinvention, and time constraints. The biggest
challenge was that the reinvented Hayden Library would re-open with only one third
of the volume count previously supported. The larger percentage of collections
managed by ASU were to be stored off-site or made available through online
platforms. McAllister realized that developing a guiding concept and purpose for
each collection area specified in the space design was key to selecting the highestpriority materials for the reinvention project. In addition, because only a percentage
of materials would re-enter Hayden Library, effects on all other collections locations
needed to be considered in the design, process, and implementation timelines. In
addition to design constraints, McAllister was also interested in refreshing the print
collections with new updated volumes. With these ideas in mind, the project team
aimed to answer the question: Within a contemporary educational environment,
where online resources are readily available, how should you go about selecting
print works for a reinvented and inclusive library space? One of our answers was
to work with our user communities to understand which types of tangible materials
needed in-person engagement, and of those, which were the most important to
have present in Hayden Library. In pursuing this vision, McAllister also wanted to
empower the project team to seek creative and novel ways to solve problems during
all phases of the project. Taken all together, this design approach was informed by
McAllister’s extensive experience as a curator and librarian working with faculty and
students within academic architecture, landscape architecture, city planning, and
design programs.

3

As of May 2024, an article manuscript describing this process, “Experimental approaches to
transforming academic library print collections,” is in submission.

Enhancing an academic library renovation project

9 of 47

Selection
Traditional good practice for selection of library materials utilizes well-crafted
collection development policies that inform selection decisions made by library
workers. Prior to the Hayden Library renovation, McAllister led the development
of an overarching collection philosophy that governs policies in focused subject
areas, such as those guiding collection development in Psychology, Government
Information, or Geospatial Information. The collection philosophy guides collection
decisions through a mission-aligned commitment to student success, excellence
and innovation in research, and social embeddedness in Arizona’s communities.
When selecting for the Hayden Library reinvention, there were many challenges in
broadly conceiving of the physical collections that re-entered the building. We had
millions of volumes across our collections from which to select, and the selections
needed to align with the building reinvention concepts set forward as part of the
project. In addition, the project team sought to utilize what we learned as part of
the eight collections experiments and other “future of print” grant-funded research
and work. The reinvention articulated a vision for contemporary teaching, learning,
and research spaces that were flexible and supported a range of activities and
technology integrations.
Using these starting points, McAllister synthesized the inputs and constraints into a
design problem and then led a design charrette with a group of library personnel to
develop user-centered collections concepts for each floor and shelving area. The
group brainstormed about possibilities for each floor, noting areas of potential and
possible challenges. Then, each person participating in the charette spent some
time individually working through their ideas by drawing on copies of each floor
plan. When satisfied with their proposed ideas, each person presented their favorite
ideas for each floor to the broader group. The group collectively discussed each
one, pulling the favorite ideas and drawings up to pin on the wall. The discussions
in the charette helped the team develop a collections narrative for each floor.
With our draft in hand, we presented these designs for discussion with library
administration, librarians and archivists, and user services staff. McAllister and Doan
also gave poster presentations at various faculty events to get input on our proposed
directions. With positive feedback and some adjustments, the final collections design
satisfied all of the criteria set forward in our design problem and still adheres very
closely to the following narrative:

•

Concourse level: This level hosts eight new University classrooms. With
shelving and soft seating areas adjacent to each classroom, the books on
the concourse level are selected in conjunction with faculty and students to
enhance the curriculum taught in these classroom spaces. This necessitates
a team approach to curation as an extension of our engagement activities.
ASU Library’s curation team aspires to work collaboratively to present

Enhancing an academic library renovation project

10 of 47

featured collections that change and grow in concert with coursework and the
activities slated for this floor.

•

•

•

First floor: The first floor of Hayden Library presents a grand entrance to the
library that is also an introduction to ASU Library’s collections. The lobby is
flanked by two grand reading rooms: the formal Wurzburger Reading Room
for distinctive collections and the Luhrs Arizona Reading Room, which is
a reading room that is open to all. The open stack collections in the Luhrs
Arizona Reading Room feature books and materials related to Arizona
and the southwest region. The collections were selected from what were
previously the closed stacks of the Greater Arizona Collection, the Chicano
Research Collection, and the Labriola National American Indian Data Center.
An interactive collections exhibit space and a small local zine collection also
invite visitors to explore the scope and variety of print collections pertaining
to Arizona.
Second floor: The second floor was designed primarily for student study,
so it is home to ASU Library’s community-curated collections, many of
which are selected by and for students. Eye-catching and browsable
collections are presented by utilizing display strategies and arrangements not
traditionally found in academic library stacks. Using print materials as tools
for engagement is intended to encourage reading and foster lifelong learning.
Students can explore topics of interest, organized by thematic categories, in
our Sun Devil Reads collection, and browse themed book displays located
next to several of our seating areas. This floor also hosts collections of
the Labriola National American Indian Data Center as part of a space that
supports enactment of ASU Library’s Indigenous land acknowledgement of
the twenty-three Native Nations that have inhabited this land for centuries,
including the Akimel O’odham (Pima) and Pee Posh (Maricopa) Indian
Communities, whose care and keeping of these lands allows us to be here
today (Arizona Board of Regents, 2024a).
Third floor: Hayden Library’s third floor was conceived of as home to a maker
space, a geospatial hub, a data science team, presentation practice spaces,
silent study for graduate students, and a workspace for visiting scholars. This
floor demonstrates the expansion of research collections far beyond the book.
The Map and Geospatial Hub provides access to thousands of cartographic
resources while supporting geospatial technologies. The Unit for Data
Science and Analytics offers numerous ways to explore and transform vast
digital collections and resources to discover new insights. The Makerspace
displays samples and specimens for exploration and to invite inquiry. Small,
focused featured collections complement the services available.

Enhancing an academic library renovation project

•

11 of 47

Fourth floor: This floor was conceived of as a space for individual study,
focused research, and small group work. The open stacks on Hayden Library’s
fourth floor are research collections designed to change and grow with our
community. They include highly used and recently acquired items of relevance
and interest to researchers in history, humanities, literature, and related
disciplines. This floor also features the Scholars Enclave, a curated set of
volumes that redefines the traditional reference collection to feature works
that are suited for in-person research. The perimeter shelves house featured
collections that highlight scholarship and offer insight into our book collections
over time.

With our draft narratives in hand, our next challenge was to engage in thoughtful
selection of the materials to include in the new Hayden Library collections.
Encompassing nearly four and a half million open stack resources across all ASU
Library locations from which to choose, the reinvention of Hayden Library allowed
our team to be curious and creative in the overall approach to collections design.

Materials and methods
In this section, the project team lays out the approach and reasoning methods
used in the selection of volumes for the Hayden Library reinvention project. We
purposefully chose to limit our selection of materials to print volumes only, continuing
to provide access to microforms, audiovisual materials, government information,
and newspapers from other library and storage locations. We transformed our
microforms services to primarily an online fulfillment model and migrated most
periodical subscriptions to online formats. We also decided to store many serials
in high-density storage, providing scanning services to fulfill article-level content
requests to users through web-based document delivery. In leveraging existing
online services, the project team limited selection of print volumes for Hayden
Library to those that would be most engaging to users of the library spaces.
Based on the floor-level collections designs and draft floor narratives, the project
team began a more detailed process of working with the architects and furniture
design teams to specify each shelving area. The team worked through various
furniture options regarding shelving type, height, configurations, and usability
issues. Using this feedback, the architecture firm developed a modified set of
plan and elevation drawings of shelving on each floor, which we used to estimate
volume counts in each area. Planning for each collection location and each shelving
section of the renovated Hayden Library was complex and required three different
approaches and methods for selection and implementation. We will describe them
in this section as 1) small featured collections, 2) a medium-sized, student-focused
collection titled “Sun Devil Reads,” and 3) a large-scale open stack research

Enhancing an academic library renovation project

12 of 47

collection. First, however, we will describe our overall approach to planning how to
use available shelving to determine the volume count for each area.
Overall approach to shelving
The shelving elevation drawings (Figure 1) depicted shelving units of varying heights
on the first, second, and fourth floors, with perimeter shelving that is interspersed
with study desk nooks and display areas. This arrangement presented several
challenges, including usability, accessibility, and ensuring that users understood how
the LC call numbers used in our collections could flow around the space and around
these desks, seating areas, and displays.

Figure 1: Sample elevation drawing of perimeter shelving

To understand how to best design collections to fit the shelving in each space,
we reviewed possible ways that our collections could “flow,” or be shelved in
logical order, throughout the new spaces to achieve a satisfying user experience.
The project team researched other libraries’ floor plans to investigate if there
was an “ideal” flow of call numbers around a library floor to maximize usability
of the collections. The team also sought to understand if and how other libraries
incorporated perimeter shelving in their floor plans and, if so, how the call numbers
flowed between shelving areas. As a benchmarking exercise, we gathered a sample
of publicly available floor plans from other libraries and modeled how our collection
might look using similar layouts and flows, noting that perimeter shelving is most
typically used only when it is parallel to closely adjacent rows, which would have
excluded nearly all of the perimeter shelving in the layout intended to house the
fourth floor stacks. In most of our test design iterations, the presence of desks,

Enhancing an academic library renovation project

13 of 47

displays, and built-in seating areas within perimeter shelving units made it difficult
to incorporate them within a flow of call numbers in a given area. The project team
determined that the continuous interruption of call numbers due to these desks,
displays, and seating areas would make it unclear to users which direction to turn
when trying to retrieve a particular book, and staff would also find it confusing to
reshelve and retrieve particular call numbers. Ultimately, this investigation led the
team to dedicate the “interrupted” sections of perimeter shelving containing desks,
benches, or displays to house featured collections, with the “uninterrupted” sections
of perimeter shelving incorporated into the collections flow.
Small featured collections
“Featured collections,” within this project context, are small collections of open stack
volumes designed to invite browsing and community participation, with the books
or themes chosen by our communities to reflect the diverse needs and interests of
users. For the Hayden Library reopening, we planned a series of featured collections
of around fifty to two hundred books each. The Curation Specialist, Emily Pattni,
collaborated with student groups and faculty to co-design browsable collections
that created learning opportunities for such topics as teaching diversity through
children’s literature, environmental awareness, and representing marginalized voices.
We encouraged collaborators to suggest book titles regardless of whether or not we
already owned them. Although it meant purchasing a large number of new books,
we saw this as an opportunity to identify gaps in our holdings and encourage our
communities to participate in the collection development process.
To order the new titles, we worked with the library’s Acquisitions team to develop
a new process for placing, purchasing, and receiving orders for physical materials
designated for featured collections. Books, DVDs, and other physical items to be
purchased were listed on a spreadsheet with necessary bibliographic data, such as
title, author, ISBN, publisher, and year of publication. Pattni then submitted a request
to purchase each title using a form, and Acquisitions staff rush-ordered these
materials through approved vendors, such as ProQuest or Amazon. Titles ordered
in this workflow were routed through a special physical processing queue that sent
the volumes to Hayden Library with work orders routing them to the staff responsible
for staging collections to move into the renovated building; this process also retained
dust jackets for Preservation Services to attach.
When books were ready for the shelf, library staff updated them to the appropriate
location code corresponding with the assigned featured collection. With over
thirty featured collections, we needed a systematic approach to display or exhibit
locations. Items in an exhibit location are assigned a temporary location code
while retaining their permanent location code within their record. Because featured
collections are intended to remain in place for long periods of time, many with no

Enhancing an academic library renovation project

14 of 47

planned end date, we needed a system that would allow staff and users to locate
books in a featured collection without an endless proliferation of temporary location
codes. Rather than request the creation of a new location code for each collection,
we designed groupings of shelving areas based on their proximity, giving these areas
names that corresponded to assigned location codes. The library’s online strategy
team adapted a practice already in use for course reserves to temporarily append
call numbers in Alma, the library’s catalog service platform; eventually, this team
created an API for staff to add and remove the names of featured collections.
For example, an item belonging to the “Utopias and Dystopias” featured collection in
its new temporary library location would have the following data in its record:

•
•

Permanent call number and location: HX810.5 .E54 2001 in Noble
Library Stacks
Temporary call number and location: HX810.5 .E54 2001 Utopias in Hayden
Library Lobby

We also worked with the library’s graphic design and communications team to
design and incorporate clear and pronounced signage that distinguished featured
collections from the continuous stacks. Consistent graphic communication, along
with displaying books with their original covers on bookstands, resulted in a series
of visually appealing featured collections that can be easily moved from one location
to another. This allows the collections teams flexibility in customizing featured
collections and adapting library spaces for events and special initiatives.
Student-focused collections - Sun Devil Reads
“Sun Devil Reads” is an open stack collection developed specifically with
undergraduate students in mind. Due to the student-study design focus of the
second floor of the Hayden Library reinvention project, the project team aimed
to present an open stack collection that would engage students and encourage
reading and exploring books beyond the classroom. We gathered feedback from
our eight collections experiments, along with our staff’s knowledge about the
interests of student communities that use library spaces, with an intentional focus on
making spaces designed to meet student preferences for study and relaxation into
welcoming areas for traditionally underserved communities. By providing books that
appeal to a wide range of self-identified areas for academic and personal growth,
we sought to equip our communities with resources to support the development of
lifelong learning and enjoyment.
Library administration approved one-time funding to create an opening day, shelfready collection. Based on the available options from our contracted vendor,
ProQuest, we chose to model our collection after the Dixon leisure reading collection
at Princeton University, which focuses on contemporary works, both fiction and

Enhancing an academic library renovation project

15 of 47

nonfiction, for general and leisure reading.4 Working with a vendor was desirable for
an “opening day” collection because they agreed to provide us with cataloged, shelfready volumes, and they could potentially provide new volumes on an annual basis.
We specified publishers, genres, subjects, interdisciplinary themes, and reading
levels, and the vendor completed the selection process on our behalf. Reviewing
sample lists and providing feedback led to the selection of 11,000 fiction and
nonfiction books published from 2017 to 2019.
Our original intent was to organize these materials following BISAC subject
headings,5 but the vendor was unable to provide classification services within
our budget, and the number of volumes involved was too great for our metadata
specialists to classify each one using a system unfamiliar to them. This led to the
decision to create our own approach to organizing books for browsing. By sifting
through sample lists, grouping together subjects and genres, and looking for
patterns, we eventually identified a list of around twenty-two browseable categories
that integrated fiction and nonfiction into thematic clusters.6 We chose category
names that invite exploration and describe a wide range of subtopics. For example,
we grouped books about food, fashion, artists, and art forms in the Lifestyles and
the Arts category. Working from a vendor-generated spreadsheet that provided
titles, authors, and subjects, we sorted the majority of the books into such categories
as Identity, Mystery, and Popular Culture prior to their arrival. There were, however,
a number of books that we left unassigned due to a lack of information. As books
began to arrive, we manually sorted these remaining books into categories and
rechecked assigned categories for other books, reassigning them as we saw fit.
To address the volume of shipments required to accomplish this project (over a
thousand books received per week for ten weeks), staff at our high density storage
facility on the Polytechnic campus handled receiving and checking in these books,
following guidance from the Acquisitions and Metadata teams. Books were then
sorted and staged based on the assigned category. These books remained on
staging shelves until they were moved into the newly renovated Hayden Library, at
which point they were installed on the second floor, shelved in LC call number order
within each category section.
Designing test-fits based on shelving capacity
In designing the layout for Sun Devil Reads, our goal was to incorporate elements
of retail bookstores that invited library visitors to browse print books and discover
something new to read among varying display layouts. We also recognized that we
4

https://library.princeton.edu/collections

5

https://bisg.org/page/bisacedition

6

For the full list, see: https://lib.asu.edu/collections/open-stacks/hayden/2nd-floor/sun-devil-reads

Enhancing an academic library renovation project

16 of 47

needed to design this section to be easily adjustable over time, as usage data will
inform the addition of a small number of books annually to keep the collection fresh
for students.
Using the floor plans provided by the architects, and with estimated counts of books
per category based on the pre-arrival category assignments, Tammy Dang, Data
Analysis Specialist, created a layout for categories to fill the shelving sections. Using
playtesting, Dang determined potential shelf fill rates and capacity. Visual mockups of
shelves allowed the project team to compare traditional shelving styles with displayonly shelving and also with blended shelving options. Eventually, we chose five style
configurations (Table 1) at different levels of aesthetic fullness to provide estimates
for ways in which categories could take up more or less space in a shelving area.

17 of 47

Enhancing an academic library renovation project

Option 1:

Option 2:

Option 3:

Option 4:

Standard Classic Shelving Shelving Split Minimal
Shelving and Display
and Display Display
Est. Volumes
Needed
19,278

13,770

Standard Shelving
This is traditional standard shelving, often
found in academic library stacks. All
books are shelved flush to the left with
spines out.

6,885

2,034

Option 5:
Display
Combo #1
6,364

Est. Volumes
Needed:
19,278

Classic Shelving and Display
This shelving mixes standard shelving with Est. Volumes
Needed:
1-3 books displayed on bookstands with
book covers facing outward.
13,770
Shelving Split and Display
This nontraditional approach encourages
browsers to scan titles on both the left
and right sides of the shelf.
Minimal Display
This shelving style uses the fewest books
to appear “full,” but it requires some
interspersed shelving with more books so
that a selection is available to replenish
the empty spaces when books are taken.
Display Combo
These mixed display options depend
on shelf height. For example, this
combination varies the Shelving Split and
Display option with Minimal Display.

Est. Volumes
Needed:
6,885

Est. Volumes
Needed:
2,034

Est. Volumes
Needed:
4,832 6,364

Table 1: Shelving style configurations and estimated volume counts for Sun Devil Reads
category sections.

Enhancing an academic library renovation project

18 of 47

Once we established capacity estimates based on the display style options, Dang
was able to work with an editable surrogate for the shelving layout to incorporate
each of the categories onto the shelving. In an example configuration (Figure 2), we
placed categories in sections based on where they fit by linear feet of space and
subject cohesion.

Figure 2: Sample category layout for Sun Devil Reading shelving.

Following a practice similar to what we developed to manage featured collections,
we included the assigned category at the end of the call number in the catalog

Enhancing an academic library renovation project

19 of 47

record.7 Within each category, books are shelved by call number, but the final layout
relies on signage to separate each group and make browsing easy for users.
Large-scale open stacks for research
To fill the newly renovated fourth floor of the Hayden Library, while also designing
a complementary stacks collection for Noble Library, the project team collaborated
with staff from throughout the organization to design and implement novel methods
for selecting and identifying books to be moved, using data to inform a largescale analysis.
The methods needed to accomplish the following:

•
•
•
•

They needed to generate a layout design for the stacks collection that
invites a logical approach to browsing sections of the collection or locating
known books.
The resulting collection should also provide a balance between
responsiveness to usage patterns or other indicators of potential interest in
works and a proportional representation of the library’s overall holdings.
The methods needed to be responsive to the varying book storage locations
in which volumes were temporarily held.
They also needed to be feasible to implement with available staff, student
workers, and contracted movers.

The staggered relocation of hundreds of thousands of volumes was a three-phase,
twelve-month, large-scale collections project, requiring the involvement of over fifty
staff members and around ninety moving truck trips. The major stages of planning
used to execute this project are outlined in the following sections.
Assessing ASU Library’s collections
The Data Analysis Specialist, Tammy Dang, began surveying the existing print
collections by creating an internal report summarizing books in each library by
collection location and comparing collections by percent of distinct loans. Out
of 1.2 million volumes remaining in public-facing library locations as of mid-2018,
59% had been borrowed at least once. Over 600,000 items had already been
moved to Noble Library from Hayden Library prior to renovation, most of which
were in the Literature and History subject areas. Through the report, the team
gained a better understanding of the content and scale of items in these public-

7

In Alma, temporarily appending a call number can only be done for holdings in temporary location
codes, so these books are assigned their permanent location as a temporary location.

Enhancing an academic library renovation project

20 of 47

facing collection locations. This information later helped us create targeted volume
counts in proportion to the collection size and usage, as well as determine which
LC subclasses would return to Hayden Library and which would find a new home at
Noble Library.
Next, the project team wanted to better understand circulating collections that fall
within the Arts and Humanities subject areas. To create an internal report for this
analysis, Dang selected the LC classification ranges of A-BT, N-NX, and P-PZ as
high-level markers, resulting in a grouping of nearly 900,000 items. This second
report focused on the loan trends of LC classification ranges by date of last
circulation and by publication year, and it separated audiovisual materials from nonaudiovisual materials. Dang also used narrower LC subject headings, down to the
subdivision level, in order to view the collection through finer lenses. Both the general
internal report and the report on the Humanities subject areas allowed the project
team to test different filters and views of the data by most recent loan date, percent
of items loaned, percent of items loaned by recent 5-year increments, subject area,
material type, publication year, year added to the collection, and assigned location.
With this data, we could highlight important factors within a volume’s loan history,
including recency of loans and proportion of recently loaned items within an LC
subclass. Later, we incorporated these factors into the selection criteria.
In a third internal report, Dang compared how selected subclasses in the Humanities
and Social Sciences varied in distinct loans within the last five years, distinct loans
within the last ten years, and loans counted in all available circulation data (going
back to 1996). We eventually determined that books that had been borrowed in the
last ten years (2009-2019) should be moved back to Hayden Library to seed the
contemporary research collection, because the difference between the number of
items borrowed in the last five years and the number borrowed in the last ten years
was negligible and the project team realized that circulation data from the past
few years (2017-2019) represented a period of exceptional disruption, with major
collection moves beginning prior to the building renovation, coupled with the library’s
transition to the Alma library service platform. Including books circulated within the
last ten years provided a more stable representation of collection usage over time.
Flowing, scaling and selecting
To determine how many books we needed to select for the initial fill of shelving, the
project team started with a shelving analysis based on the architects’ floor plans and
elevation drawings. While we knew the estimated capacity of the shelving based on
its dimensions, we agreed that we could address an important design element by
matching the collections to the visual breaks present in the space to enable easier
browsing and findability. Determining the layout and order in the flow of call numbers
was essential in calculating the final available collection size for each LC subclass.

Enhancing an academic library renovation project

21 of 47

Based on the benchmarking exercise discussed above, Dang created preliminary
sketches (Figure 3) of potential flows within the stacks.

Figure 3: E
 xamples of potential flows for putting LC call numbers in order in the fourth floor stacks.
Note that these diagrams are in reverse to the final layout.

After modeling flows using sample floor plans, we sought the expertise of library
staff to refine the floor plan. With a significant amount of trial and error, and several
rounds of feedback from stakeholder groups, including staff responsible for shelving
and frontline reference, as well as students and staff accustomed to browsing
the stacks, we agreed on the final collection flow. This design (Figure 4) takes a
counter-clockwise, or left-to-right, and head-to-tail pathway through the floor, from
the elevator entrance perspective. Four perimeter quadrants consisting of shelving
interspersed with seating and display spaces are reserved for featured collections.

Enhancing an academic library renovation project

22 of 47

Figure 4: Final flow for putting LC call numbers in order in the fourth floor stacks.

While we later made minor adjustments to start and stop points for LC subclasses,
we continued to follow the overall principles of keeping the banks cohesive
to minimize user confusion over where to look next when seeking a specific
call number.
In order to determine the total space available for the fourth floor research collection,
we combined the measurements provided for new shelving units with an estimated
average book depth. Traditional bibliographic description only provides the height
and width of books, not the depth or thickness. Fortunately, academic books tend to
fall within a fairly close range of possible thicknesses. Our initial hypothesis was that
it was most efficient to use the same conversion factor (12 books per linear foot) for
each subclass. By sampling linear feet measurements for books on existing shelves
by subclass and taking the averages, Dang gleaned a more accurate estimate of
how many volumes per linear foot of each subclass to include. We later found that in
practice, this method resulted in some miscalculations in the largest LC subclasses
(e.g., those with 20,000+ books); however, it served as an improvement on the
original 12 books per linear foot metric.
To build an engaging collection at scale, the project team wanted to select books
that were most borrowed by our users. Applying the percent of items loaned in
the last ten years to the total count of volumes for each subclass, we selected a
proportional amount in each targeted subclass to bring back to Hayden Library.
Rather than construct our selection plan solely on what we had in our collection
or what we had loaned, this allowed us to base our selection on the intersection

Enhancing an academic library renovation project

23 of 47

of existing items and relative demand for those items. Dang assigned each subclass
a weight based on the proportional usage of books in the subclass, and this weight
was then applied to the percentage of total holdings in the subclass. This allowed us
to stretch and shrink footprints based on overall holdings and current usage.
Using the new floor plans, we adjusted the proportion of footprints as needed to
create clean subclass breaks between each physical bank of shelving. With the
average number of books per linear foot by subclass conversion derived from
sampling above, Dang calculated the estimated total number of volumes needed
to fill the shelving space, leaving room for future growth within the collection. In
most cases, we sought to achieve a fill rate of around 70%; later, adjustments were
made based on the final count of books that could be selected and moved for a
given subclass.
With target counts determined, the project team needed an approach to select
which volumes would appear on the shelves. While we experimented with several
options, we had a consistent challenge finding clean and consistent data fields in
open data sources that we could connect with our bibliographic data; for example,
we were unable to effectively join our collections data with book reviews. Instead,
we eventually settled on an indexed approach, which leveraged existing, consistently
available data from Alma, such as loan history and general item record information,
to build a selection methodology that we could use and reuse at scale.
The indexed approach we ultimately selected leveraged the expertise of library
personnel who had deep experience working with the communities we serve. The
project team started by interviewing staff and librarians with varying perspectives,
experiences, and relationships with library stacks to gain insight into what is
valuable in the data. We already knew that a new book with many loans would
be preferable to an old book that has never been checked out, so we sought to
identify criteria margins and determine where thresholds and cutoffs might be. To
determine appropriate thresholds and weights for various factors, Dang conducted
a simple version of A/B testing (Table 2), offering interviewees choices between
two items from the same subclass with comparable loan statistics. Along with
their choices, they also provided reasoning behind their decision-making in each
case, as well as explanations for why they gave weight to any particular field under
certain conditions.

24 of 47

Enhancing an academic library renovation project

Choice A
LC Subclass

[GV1580-1799.4] Recreation.
Leisure—Dancing
# of Loans
11
Last Loan Date Dec 2014
Publication Year 1964
Creation Date
1996

Choice B
[GV1580-1799.4] Recreation.
Leisure—Dancing
6
May 2015
1997
1997

Table 2: Example of an A/B testing choice.

Based on our small choice experiment, we found that one of the most important
factors in every case was the most recent year an item loaned, representing the
last time an item was relevant to a user, which we categorized as the “Last Loan
Date Year.” The total number of loans per item also weighed heavily in these tests.
Ultimately, Dang created a formula to select books based on blending four criteria:

•
•
•
•

Absolute Newness - when the item was published

»

Publication Year (2019, 2018, …)

Recency of Loan - how recently our users were interested in it

»

Last Loan Date Year (2019, 2018, 2017, …)

Number of Loans - how popular it is

»

Num of Total Loans (from highest to lowest)

Relative Newness - how new it is to our collection

»

Creation Date8 (2019-01, 2018-12, 2018-11, …)

To see an applied example, consider a sample of five books, four of which had been
recently loaned and one of which had been recently published. The five books (Table
3) are ordered based on these criteria and then receive an overall rank within the set
based on the order. The bolded entry indicates the data point that best determines
the ranking for the item within the set.

8

In an example of the learning curve we encountered with this project, many of our early analyses
relying on the item’s “Creation Date” as an indicator of newness to the collection were upended
when we realized that a concurrent bibliographic remediation project conducted by the library’s
metadata team meant that many older books had brand-new records in the system.

Enhancing an academic library renovation project

Rank Publication Last Loan Number
Year
Date Year of Loans
2020
1

2

2010

2019

3

3

2018

2019

1

4

2015

2019

1

5

2010

2018

10

25 of 47

Relative
Newness

Reasoning for Ordered
Sort
New book (usually
published within the last
few years)
2010-08-01 Recent Last Loan Date
Year: 2019
2019-01-01 Same as above, but lower
in sort due to fewer total
loans
2015-07-01 Same as above, but lower
in sort due to more time
since acquisition
2011-04-01 After all books for one
Last Loan Date Year are
ranked, the books from
the previous year follow,
in order first by number
of loans and then relative
newness.

Table 3: E
 xample demonstrating how criteria are applied to order and rank books in a set.

Based on these criteria, Dang was able to calculate a numerical rating for each book.
Initially, we attempted to use approximately ten data points to establish markers in
the data to create one catch-all formula to rate individual books based on the four
criteria. Our index values needed to be of great magnitude to handle the granularity
we required between layers of criteria, so this method was ultimately unsuccessful.
Thus, in the interest of time, we ultimately used a stepwise formula that incorporated
the four quantitative factors for each LC subclass, which provided the necessary
granularity to achieve a target list. The exact cutoff values for each criteria expanded
and contracted on a subclass-by-subclass basis. Therefore, these values fluctuated
to fit their designated volume targets and subclass characteristics. For example, we
might expect an item in subclass BF Psychology would have a different Publication
Year cutoff than an item in PN Literature.
Sample criteria for a single subclass
A sample output of stepwise criteria (Table 4) demonstrates how Dang generated
a balance of Number of Loans to Last Loan Date Year, thereby allowing items that

26 of 47

Enhancing an academic library renovation project

were more recently added to the collection to have sufficient time on browsable,
public-facing shelves. In this example, which follows these adjusted stepwise criteria,
items with any Creation Date from 2016 or later and Publication Year from 2000 or
later were selected. Items with a Last Loan Date of 2020 and a Creation Date of
2010 or later required two or more loans to be selected; if the item’s Last Loaned
Date was in 2020 and the Creation Date was earlier than 2010, then it would require
three or more total loans in order to be selected. This stepwise pattern repeats for
every Last Loan Date Year, where earlier years require higher numbers of loans for
items to be selected.
Last Loan Date
Year
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
any year
any year

Num of Loans

Creation Date

Publication Year
2000+

3+
2+
3+
2+
3+
2+
4+
3+
5+
4+
6+
5+
7+
6+
8+
7+
9+
8+
10+
9+
11+
10+
0+
1+

2016+
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
2010 +
any year
any year

2015+
2014+

Table 4: Sample output of stepwise criteria. These criteria were adjusted for each LC subclass to
produce the approximate count of items corresponding to the estimated capacity of the desired
location within the overall layout.

Enhancing an academic library renovation project

27 of 47

Our selection process did not solely abide by this ordered criteria methodology.
Other factors considered included topics of local and regional interest (e.g., Arizona
and the Southwest) and the language of the work. We also manually reviewed multivolume sets when different volumes received different rankings based on usage or
publication date.
Selection in practice
After establishing how many books of each subclass we needed and determining
how to select the “best” books based on ordered criteria, it was necessary to
identify efficient methods for physically tagging or removing these books from their
current locations and preparing them to move into Hayden Library.
There were several interrelated challenges involved in addressing this question. As
discussed above, while some books remained in open shelving on Tempe campus,
others were transferred to Polytechnic campus, and some of these subclasses had
since been fully accessioned into the high-density storage facility, while others were
packed tightly (and not always in order) in closed temporary shelving. The realities
of library shelving practices over time further complicate the selection process.
Books accessioned into a high-density storage facility are extremely likely to be
exactly where they are expected to be, as they have a unique storage location
identifier added to the item record that pinpoints a location in a specific tray on one
of thousands of shelves. Books in other shelving systems, however, may or may
not be present, as they may be missing, mis-shelved, or on the shelf but lacking a
catalog record.
In our selection practices, we also needed to accommodate the fact that books that
are accessioned into high-density storage can only be selected with an item-by-item
pick list. Library staff developed a high density deaccession (or “de-processing”)
workflow in which books were pulled from trays, storage location identifiers were
transferred to incoming books of the same size and thickness, and the selected
books were staged in call number order for moving to public-facing shelving.
Because of the intensive labor involved in this process, we had to limit selection to
the top-ranked “best” books for subclasses that were already fully accessioned
into the storage facility. At the same time, a pick list would clearly not be effective
for books held in open shelving. Although some sections had long ranges of “best”
books, they were often scattered all throughout the shelves. It was clear from the
start that we did not have the staffing to individually flag or pull every “best” item
from the list (it likely would have taken more than 3,000 hours for staff to pull
120,000 books).
Dang ran a series of trials using Alma and Alma Analytics to see if it would be
possible to determine exactly what is in a given section of books, down to the LC
subclass or even by LC call number. If so, it would be possible to select books from

Enhancing an academic library renovation project

28 of 47

a collection on a shelf-by-shelf basis. However, this method would require staff to
record all start and stop barcodes for each range in the stacks; after conducting
trials of this method, we determined that the potential for error, especially in misshelved books at the end of ranges, would prove too costly and inaccurate for the
meticulous effort needed to get all the data.
Needing another approach, we conducted a series of four informal interviews with
library staff who had significant prior experience in preparing and conducting book
moves. Based on the discussion in these interviews, we settled on a “chunking”
approach, where sections of books on shelves would be flagged to move. This
approach requires fewer books to be separately identified and flagged. Moreover, we
considered these “chunks” as a way to preserve some browsing serendipity, when a
book that is shelved next to a sought-after book may be of interest to a browser.9
“Cascading Windows of Sets”
The technical challenge then became programmatically identifying sections where
the most “best” books reside in clusters close to one another, and then select the
best sections based on available space. Dang consulted with Professor Phillip Miller
from ASU’s School of Computing, Informatics, and Decision Systems Engineering,
who advised us on the logic and efficiency of our algorithm. With his help, Dang
wrote an algorithm in Java using the free Eclipse tool.
The algorithm needed to provide a simple solution to an optimization problem: how
to pull as many “best” books as possible using the fewest instances of physical item
searches within the shelves while under the constraints of limited programming
sophistication and time. Dang developed a workable, naive10 algorithm that found
concentrations of “best” books on the shelves. Dubbed “Cascading Windows of
Sets,” the resulting algorithm identified sets of the highest-indexed books within
a given population, while allowing for the adjustment of acceptable gaps of nonindexed books to include in an output “chunk” (section of books in order on a shelf)
to achieve a desired target count of items with a locally minimized count of items that
would be designated as the “start” or “stop” for each section that needed to move.

9

One downside of the “chunking” approach turned out to be the lack of precision in updating
locations for selected books; while we made significant efforts to maintain a parallel listing of
flagged groups of books and item records in call number order, shelving mistakes and errors
ultimately meant that some books were incorrectly updated (or not updated). Post-move, we
continue to remediate these issues through shelf-reading and other cleanup projects.

10 A naive algorithm relies on computing many simple calculations rather than a few sophisticated
ones. Since the computing power required for simple calculations is minimal, the algorithm takes
relatively little time to run even with millions of lines of input data.

Enhancing an academic library renovation project

29 of 47

Results
Preparing books to move
In late January 2019, we coordinated the movement of around 125,000 open stack
volumes from temporary shelving in the high-density storage facility to another
nearby temporary storage facility to begin to stage books for Hayden Library. At
the same time, we began to plan in earnest for the collection moves needed to
populate Hayden Library with a collection synthesized from all the selected volumes
at all of the various storage locations. We created a method to identify the books on
public shelving that we had selected for Hayden by printing colored flags to show
the contracted movers where each “chunk” started or stopped. Each flag had a call
number, a direction to either “start” or “stop,” and a note for library users to inform
them about the upcoming book moves.

Figure 5: Sample of start and stop flags

Dang developed a tool in Java to print these flags, and teams of student workers
placed each flag inside of the correct book. As flags were matched to volumes
and inserted, adjustments were made and recorded to account for missing books,
monographic series, and other problems as they arose. Immediately prior to the
move, sets of selected items (or the non-selected items between selected sets,
depending on the final destination of each segment of the move) were flipped spinedown so contracted movers knew which books to take and which to leave. Once
groups of books designated for off-site storage were moved from the shelves, staff
and contracted movers condensed the remaining books.

Enhancing an academic library renovation project

30 of 47

Preparing destinations
To prepare to conduct the move, Dang created a Percent Fill spreadsheet that
produced estimates for how many books per subclass would fit on the new Hayden
shelves. This gave the team a quick method for determining how full to fill the shelves
in each LC subclass section, resulting in tape marks added to the shelving to guide
the contracted movers in loading books onto the shelves, with the goal to reduce the
amount of shifting projects staff would need to complete after the move.
The first task was to identify the smallest reasonable locator unit for the shelving
in order to calculate all fill permutations permitted within the space available. Dang
developed a system dubbed “BAUS” for Bank, Aisle, Unit, and Shelf. At the shelving
bank level, there are a varying number of aisles (single-faced rows of shelving),
and there are a varying number of units (single-faced units of shelves) within each
aisle. Units themselves vary in height and therefore in number of shelves, and to
complicate matters further, some shelving varied in width along the perimeters. The
smallest unit of shelf measure thus had to be at the level of individual shelves, each
of which is identified by its bank, aisle, and unit number.
We used this BAUS system to precisely identify anticipated starting points, stopping
points, and skipping points for each incoming LC subclass section. In an example
of the floor map (Figure 6), the area highlighted by a yellow box represents the units
of shelving at Bank03 Aisle01. In this particular aisle, there are 10 units, and each
unit is 4 shelves in height. When facing the shelves, the far-left, bottom-most shelf is
denoted as B03-A01-U01-S01.

Figure 6: E
 xample of floor map indicating the destination locations for LC subclass sections.

Enhancing an academic library renovation project

31 of 47

The project team made a custom measurement tool, similar to a yard stick, that had
marks for lengths that match various percent fills (15%, 25%, …, 80%, 95%) for the
standard shelf width. Most shelves in the new building are 35 inches wide (although
some perimeter shelves are wider), so each tick mark on the measuring stick
corresponds to the percent fill of a 35-inch shelf (e.g., 50% is marked at 17.5 inches
on the stick, 60% is marked at 21 inches, and so on).
For some sections, we were able to measure the exact linear feet of the incoming
subclass. In these cases, we simply applied the measuring method outlined above. If
the books were sitting on accessible temporary shelving, we used the method above
to quickly measure the shelves to get an accurate linear feet count, by counting
shelves based on their approximated percent fill and then using this data to calculate
the linear feet. For collections for which we only had the estimated number of books
coming in, we used the previously developed books per linear foot estimates to
calculate estimated total linear feet.
Once we had a linear foot estimate, the next step was to distribute the books among
the designated shelves. We used a calculation stored in a spreadsheet to calculate
shelf fill:

•
•
•

Top Shelves LF Actual = (Top Shelves LF Capacity) / (Top Fill % [manually
select desired fill])
Remaining LF = (Total LF) - (Top Shelves LF Actual)
Bottom % Fill = (Remaining LF) / (Bottom Shelf LF Capacity)

In some estimated linear feet cases, the books per linear feet conversion rate
produced too few books for some subclasses and too many for others.11 The
estimation also did not account for multivolume sets, which created overfill as we
sought to avoid splitting them across multiple locations.12
For the bulk of items moving back into Hayden Library, our preparations required
staging final destinations and posting moving signage. Using painter’s tape and the
measuring stick, library staff also added a “stop” point to each empty shelf based
on the anticipated fill rate and taped an “X” on shelves that should be skipped.
Moreover, during an initial walkthrough, we discovered missing or partially installed
11 Our calculation for average books per linear foot measurement by subclass was based on
samples conducted prior to the selection process and then used to estimate the width of
selected items, meaning that older and less-popular books were overrepresented in the
conversion when compared with the materials selected for the stacks.
12 When the contracted movers reached the point of loading items onto the new shelves, library
staff kept an eye on how the fill rate was proceeding, making adjustments or authorized
overflows as needed. In the weeks following the move, staff and student workers conducted a
series of collection shifts to better settle collections into their new spaces.

Enhancing an academic library renovation project

32 of 47

shelving that offset our calculations. We documented these inconsistencies,
reported back to the contractors for fixes, and readjusted our move maps for
exceptions. After conducting a final inspection, completing temporary signage, and
adding shelf fill tape markers to each shelf, we were ready for collections to move
into their new spaces.
Moving phases and process
By the time the move started in late 2019, books were prepared, flagged, or staged
at various libraries and storage locations, based on the selection criteria and
selection algorithm, for their return to Hayden Library. In all, there were three primary
processes developed and used:

•
•

•

Books that had been temporarily relocated to the Polytechnic campus as part
of a final preparatory move in January 2019 were identified and staged based
on pick lists. The remaining volumes were queued for processing into highdensity storage.
Books that had already been processed into high-density storage, including
those from earlier moves, were deaccessioned one by one from storage, using
a workflow that matched the “de-processed” book with another book of the
same size to fill the slot in the storage tray. These books were also staged in
temporary space at Polytechnic campus, along with newly acquired volumes
that fell outside the call number ranges shelved at Noble Library.
The “chunking” methodology was used to prepare to move volumes
remaining on the shelves in temporary locations in Noble Library and Hayden
Lower Level.

Beginning in late August 2019, an eight-week project relocated around 229,000
open stack collection books from Noble Library to temporary offsite storage on
the Polytechnic campus for eventual accession into high-density storage. This
move was the first to use the “chunk” selection methodology described above,
with sets selected, flagged, and flipped to aid the movers in pulling what needed to
be removed.
An intensive five-week project starting in late November 2019 brought around
187,000 books into the Hayden Library tower as part of the final weeks of postconstruction, pre-opening activity. Staged books were first moved into Hayden
Library from temporary storage on the Polytechnic campus, then flagged chunks
were pulled and moved from Noble Library and Hayden Lower Level, and then Sun
Devil Reads was moved from its staging location. This move was largely completed
in early January 2020. Staff also relocated and shelved dozens of newly created
featured collections, leaving designated spaces for additional collections slated to
be acquired in 2020, and spent many hours shifting, arranging, and labeling shelves

Enhancing an academic library renovation project

33 of 47

throughout the building in preparation for the grand opening celebration, scheduled
for March 25, 2020.13
The third phase of the move completed the project, sending around 170,000
volumes from Noble Library and Hayden Lower Level to the Polytechnic campus
for processing into high-density storage, arranging Social Sciences and STEM call
numbers (G-L, Q-T, Z) on Noble Library’s third floor, and fully emptying the second
floor of Noble Library so its deteriorating electronic compact shelving could later
be removed. This move began at the start of March 2020, was halted when ASU
temporarily ramped down onsite operations, and then resumed in June for final
completion in August 2020.
Throughout these efforts, including during the height of move activity in December
2019 and the temporary building closures from March through August 2020, fulfilling
patron print item requests remained a high priority. All of the moves incorporated a
process to create lists of barcodes that were projected to be moved at a particular
point in time, sometimes down to the day, so the location codes could be updated
on a consistent basis using a bulk update process in Alma. Fulfillment staff devised
ways of communicating about and checking for volumes that could be at one
of three possible locations, and all front-line staff provided additional guidance
to patrons on expectations for requesting books and resolved issues as they
were raised.
When the building reopened and full onsite services resumed in August 2020,
staff and student workers began a months-long project to shelf-read the newly
installed Hayden Library and Noble Library stacks collections, reconciling holdings
statements with what appeared on the shelves. Staff at the high-density storage
facility also continued the long-term project of accessioning volumes from temporary
shelving into storage, eventually completing this work in 2022.
With the completion of this major collections design and move project, the Hayden
Library fourth floor stacks and complementary collections at Noble Library became
a proportional representation of the entire collection, serving as a starting point for
growth and development, while the Hayden Library featured collections can be built
up, rotated, and repurposed based on usage data.

13 While the grand opening celebration was indefinitely postponed and eventually canceled
following ASU’s temporary building closures around March 18, 2020, ASU Library never fully
suspended its print collection services. Fulfillment of patron physical item requests continued
during the summer of 2020, and document delivery services resumed along with faculty and
student building access in August 2020. Physical item ordering also resumed, and the library
eventually acquired all of the featured collections selected for the grand opening.

Enhancing an academic library renovation project

34 of 47

Discussion
The idea of transforming collections to center our users in academic library spaces
opens a pathway for library leaders and collections managers to rethink traditional
methodologies for the selection and development of collections, including how
collections are displayed in library spaces (McAllister & Laster, 2021). Traditional
methodologies are grounded in the perspectives and experiences of the people who
make decisions about collections, with library practices and methods as an extension
of these ways of being and knowing. The design of library collections, then, starts
with the people who are making decisions, which is why it is important to have a
diverse team and to develop a safe culture within the team’s working environment
(McAllister, 2021). The approaches to collection selection and development, including
the ways in which data are used for collections design, data analyses, and operations
workflows, all flow from the world views and experiences of the team. Designing
collections requires decision-makers to consider the nature and purpose of print
collections within a contemporary academic library context, including library spaces.
The Andrew W. Mellon Foundation funding for the three-year project period afforded
ASU Library the creative freedom to explore new ideas and methods for the “future
of print” as an engagement tool within academic libraries. This project gave our team
time to be creative and generate new ways to think about collections methodologies
and processes. Traditional processes often used to manage collections in academic
libraries may need to be refactored to meet contemporary library needs and user
expectations. Creating and implementing innovative approaches to collection
management requires a high degree of personnel time to develop new collaborative
processes and operational procedures. The support for creative exploration resulted
in print collections for Hayden Library that are more inclusive and communityinformed than those resulting from prior practice.
Selection and development of collections
Traditional library selection processes typically involve professional librarians
working within the scope of their collection development policies to create
parameters for selecting materials for a collection, often using approval plans from
vendors that result in automatic purchase and shipment of books that meet those
parameters. When the books arrive, library workers generally route these volumes
to a cataloger or cataloging team. Once cataloged, the volume is assigned to a
shelf location, stickered and stamped with its classification number and property
markings, and then shelved for circulation by library workers in its designated
location. Generally, books are not relocated once shelved unless they are pulled off
the shelf for use or check-out by patrons or are removed as part of a stacks weeding
project. The selection processes we developed for the small browsing collections
at Hayden Library more closely mirror community participatory design work for

Enhancing an academic library renovation project

35 of 47

neighborhood and housing complexes that arose in the United States in the 1990s.
Architects, designers, urban planners, and activists collaborated with residents using
an action research approach, which included the people who lived in the housing
or neighborhood project area in decision-making about the development of their
environment as a mechanism for social change (Day & Parnell, 2003, p. xi). Using
community participatory design for library collections brings collections users into
the selection space typically reserved for professional librarians, serving to shift
power relations within library disciplinary practice.
By opening up traditionally cordoned-off professional spaces, library workers who
are not professional librarians or archivists are also empowered to contribute to
selection processes. In developing browsing collections, our team worked with
community members, students, library staff, and faculty using selection processes
that were more community-friendly than traditional vendor-based processes. A
variety of list-making strategies emerged from these collaborations that were not
necessarily based on the library’s current set of books available to use for our
reimagined Hayden Library collections. The community and library teams codeveloped lists of print books using a variety of means, and in many cases the library
purchased titles identified and selected for these small browsing collections, which
contributed to the richness of the library’s holdings. Removing the requirement that
collaborators already know how to use the library’s discovery tool to identify books
to include in a collection further reduces barriers to collaboration. Easy-to-use listmaking tools, such as short web forms, were the foundation of our action research
approach to include more people in the design of collections, moving library practice
toward a more social and open space of possibility.
Displaying collections in libraries
While the selection approach to the Sun Devil Reads collection used a vendorbased approval plan, the project team designed the organization and display of
this collection to be people-friendly. We spent time at the beginning of the project
researching spaces where people encounter books, including bookstores, airports,
public libraries, school libraries, and take-one-leave-one shelves in community
spaces. We developed a plan to present this collection as a departure from the LC
classification system, instead organizing the collection by user-friendly subjects,
more similar to the ways that bookstores present books to potential buyers than
to traditional academic library practices. We wanted the collection to be more
easily browseable for students, with the goal that books in this collection would be
attractive and highly used.
This collection’s design is an example of movement toward a more “whole person”
experience within the library, much like the trend in bookstores to open up retail
choices for shoppers. In 1999, Alexandra Jardine suggested that emerging

Enhancing an academic library renovation project

36 of 47

megastores, such as Borders and Waterstones, “offer the consumer much more
than just books for sale” (Jardine, 1999). These bookstores co-mingled books with
merchandise, and added coffee shops and restaurants within their walls, aiming to
create a more rich customer experience. In 2013, the manager of the famous Trident
Booksellers & Café in Boston stated that they added food, event space, and sideline
merchandise to their book offerings because “people come here for the experience,”
meaning that bookstores sought to offer an experience for the mind and the body
(Rosen, 2013). Like bookstores, libraries have changed over the last few decades
to incorporate more spaces for student services, including tutoring and writing
support, cafés, meditation rooms, and a variety of study spaces. Section by section,
the academic library can become not just a warehouse for books, but a place of
experiences created through intentional user-centered design.
Analyzing and using data for collections design
Creative design processes played a significant part in the project team’s approach
to analyzing and using collections data. The Data Analysis Specialist pursued new
ideas and approaches with a technique of quickly mocking up data in Google Slides
(as seen in Figure 3, above) and Google Sheets (as seen in Figure 6, above). The
resulting diagrams served as effective communication tools for the project team
members and other stakeholders, and they were easily modifiable to support lowstakes experimentation and change without the need to adopt and quickly grow
proficient in a graphical design program. This iterative approach allowed the team
to come to consensus more quickly, and it limited the intermediary steps needed
between the generation of an idea and a more in-depth discussion of its implications.
Renovating operations and workflows
This project resulted in many process changes within our collections work,
including such operational changes as measuring and adjusting shelving fill rates for
collection shifts of varying size. Our team found that creating documentation before,
during, and after a project is useful and valuable. As the project progressed, we
experimented and worked through design problems, discussing these experiences
with colleagues across the library. This feedback process and subsequent time for
reflection helped create solutions that fit design problem parameters and, ultimately,
communicate decisions and carry out the implementation. As an example, the
Curation Specialist developed a Featured Collections Toolkit to outline the process
for engagement and outreach to community groups and users, the logistics in
curating and maintaining these collections, and best practices in partnering with
internal and external stakeholders to communicate a collection that will engage our
users. Since the project’s conclusion, we have found that reusing the tools that we
developed for this project and retrofitting them for future uses is extremely helpful

Enhancing an academic library renovation project

37 of 47

in response to personnel turnover and for the purpose of training staff not already
familiar with these collections processes.
The Hayden Library reinvention project was completed in March of 2020, just as
the COVID-19 pandemic caused university campus closures. Pandemic-related
closures and service changes impacted library usage that, in turn, stalled our
analysis of user engagement with the newly placed Hayden Library collections. Over
the course of the project, we learned that keeping an open mind and being flexible
was important to surviving a massive library project of this kind. Author Vickie Mix
expressed this well in a description of a collections renovation project: “Flexibility
is key to surviving the stress created by changing priorities and directions” (Mix,
2012, p.156). Among many profoundly unprecedented challenges, the pandemic
caused a major disruption in our plans to evaluate the collections after installation,
due to the temporary transition to online operations and subsequent limited facilities
access. Our project team instead completed an analysis on preliminary data for the
Classroom Collections, combining loan history with user enrollment data for a set of
collections that had been launched in Fall 2019 and made available for one semester
in a section of Hayden Library, which had opened earlier than the rest of the building.
While this exploration suggested a framework for future data analysis, its limited size
and scope prevented the team from drawing conclusions that could be extrapolated
to inform future collections decisions. We look forward to further analysis projects
and hope to share this work to inform the processes of other research libraries
seeking to reinvent their print collections.
The New Open Stacks
The use of print collections in academic libraries is changing. The “future of print”
project demonstrated that print volumes can be a vital part of contemporary
academic library spaces and services. As an implementation of some of the ideas
initially set forward in ASU Library’s 2017 white paper, the Hayden Library reinvention
may also serve as an example for how open stacks can be positioned within 21st
century library space design (ASU Library, 2017).
Innovation serves to move the discipline of librarianship toward more contemporary
practice, yet removing books from 21st century library designs comes at a cost
(McAllister & Laster, 2018a). When facing a renovation project in this era, when user
study and workspaces, maker spaces, student support services, and collaborative
spaces are often preferred over collections spaces, it is imperative that collections
within library spaces are user-centered and consider the whole person in their
design. To achieve this, collections design decision-making should be centered
within a diverse team that is supported and empowered to make independent
judgements and work in tandem with library community users.

Enhancing an academic library renovation project

38 of 47

Acknowledgements
The authors would like to acknowledge all of the many ASU Library personnel who
contributed to this project and to the development of this report; an incomplete
list includes Amy Carolyn Watson, Becky Uhl, Betsy Redman, Corie Cisco, Cosmo
Anaya, Daniel Rogers, Daphne Gill, David Kellen, Heather Mossberger, Kelsey
Hinesley, Kevin Feeken, Jaime Jones, Jason Johnson, Marilyn Murphy, Mark
McCann, Mark Prestegard, Matt Ogborn, Meaghan Ward, Michelle Theodorou,
Rachel Leket-Mor, Sam Mijal, Shari Neal, Stephen Arougheti, Suzy Morgan and
Thomas Wills. The authors would also like to express our gratitude to The Andrew
W. Mellon Foundation, and to Don Waters and Patricia Hswe for their support of the
work leading to this paper.

Enhancing an academic library renovation project

39 of 47

References
Arizona Board of Regents. (2024). Hayden library reinvention. Business and Finance,
Arizona State University. https://cfo.asu.edu/hayden-reinvention
Arizona Board of Regents. (2024a). Indigenous land acknowledgement. ASU Library.
https://lib.asu.edu/indigenous-land-acknowledgement
Arizona Board of Regents. (2024b). Hayden history. Hayden 2020. Arizona State
University. https://lib.asu.edu/Hayden2020/past
ASU Library. (2021). ASU Library releases white paper
on the future of print. Library Channel. ASU Library.
https://lib.asu.edu/news/asu-library-releases-white-paper-future-print
ASU Library. (2017a). Future of the academic library print collection: A space for
engagement. Arizona State University. https://hdl.handle.net/2286/R.I.50125
Bennett, S. (2003). Libraries designed for learning. Council on Library and Information
Resources. Washington, D.C. https://www.clir.org/pubs/reports/pub122/
Crow, M. M. & Dabars, W. B. (2015). Designing the new American university. Johns
Hopkins University Press.
Crow, M. M. & Dabars, W. B. (2020). The fifth wave: the evolution of American higher
education. Johns Hopkins University Press.
Jackson, P. L. (2009). Getting design right: A systems approach. CRC Press.
Jardine, A. (1999, Nov 04). The bookstore fights back. Marketing, 28-29.
McAllister, L. (2021). “The new open stacks.” In L. McAllister & S. Laster (Eds.),
Transforming print: Collection development and management for our connected
future (pp. 3-20). ALA Editions. https://hdl.handle.net/2286/R.2.N.160096
McAllister, L. & Laster, S. (2018). “The future of print in open stacks.” In L. Meyer
(Ed.), Charleston voices: Perspectives from the 2017 conference. Against the
Grain. https://doi.org/10.3998/mpub.11281794
McAllister, L. & Laster, S. (2018a). Open stacks in library design. Portal, 18(3).
http://hdl.handle.net/2286/R.I.49681
McAllister, L. & Laster, S. (Eds.). (2021). Transforming print: Collection development
and management for our connected future. ALA Editions.
Mix, V. (2012). Renovation and roadblocks while protecting the collection. Collection
Building, 31(4), 153–57. https://doi.org/10.1108/01604951211274061

Enhancing an academic library renovation project

40 of 47

O’Donnell, J. (2017). The future of print in open stacks: A proposal. Charleston
Conference. https://www.youtube.com/watch?v=Go4GJCYCLqk
Day, C., & Parnell, R. (2003). Consensus design: Socially inclusive process. Routledge.
Rosen, J. (2013, March 10). Is the time right for true bookstore cafes? Blended retail.
Publishers Weekly. https://www.publishersweekly.com/pw/by-topic/industrynews/bookselling/article/56288-is-the-time-right-for-true-bookstore-cafes.html
Sandy, J. H., Krishnamurthy, M., & Scalfani, V. F. (2014). Repurposing space in a
Science and Engineering Library: Considerations for a successful outcome. The
Journal of Academic Librarianship, 40, 388-393.
Somerville, M. M., & Brown-Sica, M. (2011). Library space planning: A participatory
action research approach. The Electronic Library, 29(5), 669-681.
https://doi.org/10.1108/02640471111177099
Yi, Z. (2015). Knowledge management for library building design. Library
Management, 37(1/2), 2-12. https://doi.org/10.1108/LM-06-2015-0034

Enhancing an academic library renovation project

41 of 47

Appendix
The concept of the “Cascading Windows of Sets” algorithm can be better explained
using the following simplified example.
Suppose we have ten items on the bookshelf in perfect LC Classification order.
The items marked in yellow are the ones that scored highly using our selection
methodology as described in the paper above. Each item has a set position, for
which we assign an index number from 0 through 9.

Suppose we wanted to look at every possible consecutive set of three items. The
first three items on the shelves would be from index 0 through index 2, the second
set of three consecutive items would be from index 1 through index 3, and so on.

Enhancing an academic library renovation project

42 of 47

Each possible set of three consecutive items will have a summed score. We then
select the highest-scoring set from all possible choices.
In this example, Set 2 would likely have the highest score. Set 3 (index 2 through 4)
may have the same score, but for simplicity we select the first indexed set in cases
of matching set sums.

Enhancing an academic library renovation project

43 of 47

The selected indexes are then removed from the list of available set options. We
repeat this process, selecting the top-scoring set and removing those indexes
from the available pool, until we reach the total number of books needed for a
given subclass.
In a second iteration of the algorithm, we sought to increase the concentration of
target items from the selected sets by removing leading or trailing non-target items
(or lower-scoring items). Continuing this example, both sets have leading non-target
items (shown below in gray).

Enhancing an academic library renovation project

44 of 47

We can remove leading and trailing non-target items from their sets.

In an effort to increase efficiency and to reduce the number of physical items
needing flagging, we then look at only our selected sets and consider the numbers
of books in between each set, referred to as “gaps.”
In this example, we join together any selected set separated by two or fewer items
(i.e., having a gap size of two or less). In practice, the gap size is a variable.

Enhancing an academic library renovation project

45 of 47

We have now selected five items, three target items and two non-target items, while
requiring two physical items to be flagged (one for the start of the set at index 2 and
one for the end of the set at index 6).

Now suppose we were to consider every possible set of five consecutive items
instead of three consecutive items. Similar to sets of three, the first five items
would represent set 1, and the sets of five would cascade through all possible items
like before.

Enhancing an academic library renovation project

46 of 47

Without having to trim leading and trailing non-target items or adjoin sets with small
gaps of separation, we are able to produce the same highest-score set of five items
(index 2 through 6).

Enhancing an academic library renovation project

47 of 47

Our next task is then to find the optimal values for set sizes (s), gaps sizes (g),
and single selections (ss) by comparing the efficiency of combinations of these
variables. We consider the percent of target-items being selected (Percent of Single
Select) compared to the number of physical item searches for flagging that would
need to be conducted (Number of Flags). Combinations landing high on the y-axis
(gathering more of our desired items) and left on the x-axis (requiring fewer physical
item searches from the shelves) are our ideal choices for selection. Much like the
selection protocol, the combination of variables for cascading sets varied for each
LC subclass.