Report (December 2011)

If you want to comment or contribute to this wiki use the DISCUSSION tab or feel free to contact
Ken Chad ken@kenchadconsulting.com Tel: +44 (0)7788 727 845 Twitter: @kenchad



The final report is available online below. It is also available here as a PDF download


A presentation on the project was made by Ken Chad at the 'ebooks unbound' conference in Glasgow in October 2011




Patron Driven Acquisitions (PDA) and the role of metadata in the discovery, selection and acquisition of ebooks

1. Introduction

This project investigated end user’s motivation for selecting an e-book and the role that metadata plays in the discovery, selection and eventual acquisition process. The study also explores factors such as reading lists, tutor recommendations, student recommendations and other drivers that may arise through the course of a student’s and academic’s work and research. It does this in the context of the Patron (or Demand) Drive Acquisitions (PDA /DDA) model. In looking at ebooks, the study does not specifically address the issue of textbooks. Publishers remain reluctant to make textbooks available as ebooks and, in any case, PDA regimes in Higher Education tend not to be designed for or directed at the selection and acquisition of e-textbooks. The JISC ebook observatory project[i] has dealt in some depth with issues around e-textbooks. JISC has also made a collection of around 3,000 ebooks available to Further Education (FE) institutions on a national basis. [ii]

In commissioning this study the Jisc felt that: ‘A lack of understanding on the motivation for the discovery, selection and acquisition that underpin the demand-driven acquisition of e-books means that librarians and publishers are unable to identify the role that metadata plays, nor ensuring that it is available in appropriate forms at each stage of the process.’ This start makes a start on building an evidence base to assist in achieving that aim.

2. Key recommendations for Jisc (see section 8)

2.1 The market

Jisc can support the better operation of the market by facilitating the exchange of information. The case studies included in this project are a part of this information exchange.

2.2 Metadata

Jisc should use its influence in bodies such as NISO, and in particular the Open Discovery initiative to help create the most effective and efficient environment for ebooks. In addition there is opportunity for the Jisc to coordinate efforts with organisations such as UKSG (a body encompassing all stakeholders) to clarify the mechanics of how ebook metadata might best be managed and the enhancement needed by discovery services to enable the best implementation of business models such as PDA.

2.2.1 Identifiers

The growing importance of identifiers for ebook is recognised within the industry. We have already mentioned the and Jisc should engage with this group to achieve with BIC metadata futures group to get the best possible outcome for UK Higher Education

2.2.2 Social metadata

There are a number of existing initiatives which Jisc can contribute to including the NISO work on ‘E-Book Annotation Sharing and Social Reading’.

3. Project Team**[iii]**

Ken Chad of Ken Chad Consulting Ltd led the project and authored this report. Carol Thomas, formerly Head of Learning Support Services, at the University of Chester was invited to join the project team mid way through the project to assist Ken Chad Consulting with specific project tasks including the workshops, interviews and focus group.
Other members of the team are:-
  • Terry Bucknell, Electronic Resources Manager, University of Liverpool Library
  • Sarah Thompson, Content Acquisition Librarian at the University of York Library,
  • Nadine Prowse, Vice President, Operations Management for Ingram Digital's Institutional Solutions.
  • Jude Norris, Marketing & Technology Director Dawson Books Ltd
  • Scott Wasinger, Senior Director of Sales for eBooks and Audiobooks at EBSCO Publishing

4. Methodology

The project was modestly resourced project and was undertaken over a relatively short timescale. Our methodology has therefore been to involve a diversity of stakeholders (librarians, ebook platform vendors, publishers, academics and students) through workshops, interviews, focus group and case studies. Using this multi-strand approach we believe we have been better able to uncover consensus on the key issues. The project ran over the summer period in 2011 which constrained our ability to get input from some stakeholders (notably students). An extension of the project has enabled us to add more input from end users in the form of an online survey, interviews and a focus group.
In summary the following stands are referred to in this report:
  • Workshops: including the 'Jobs-to-be-done' approach
  • Stakeholder interviews
    • Librarians
    • Publishers
    • Vendors/Platform providers
    • Academics
    • Students
  • Case studies
  • User Survey
  • End user focus group
  • Other relevant work

The detail and outputs of each strand are available from the online project wiki[iv] and may be freely downloaded and re-used

5. PDA in context

Ebooks are part of the wider trend towards digital content. Furthermore what we think of as an ‘ebook’ will evolve. For example Faber have, with Touchpress, issued a version of TS Eliot’s ‘Wasteland’ which is marketed as an ‘app’ rather than an ebook.[v] The move to digital content has already gone a very long way with journals and one librarian has said: 'ultimately, we know what our users want in ebooks: the same freedom they have with electronic journals.'[vi] From a HE perspective, the 2011 Horizon report[vii] cited ebooks as one of the near term ‘technologies to watch.’ Digitisation has created content that is, potentially at least, easier to disseminate, easier to discover and easier to consume. One of the publishers we interviewed pointed at that ebooks can offer ‘functionality that you can’t provide in print: search inside, annotations, bookmarks etc’. Moreover a print book can only be in one place at a time and the print format itself puts a very effective brake on how a book is used. It is hard for it to be used simultaneously for example. This contrasts starkly with the 'non rival'[viii] nature of digital content, which in turn creates a serious challenge for publishers and others to devise business models for a new environment. For example most major ‘trade’ publishers refuse to allow their ebooks to be loaned through public libraries. ‘As more of its media morphs into digital forms, the publishing industry is undergoing a shift very similar to the one that took place in the music industry in the last decade. New business models and methods of distribution are appearing as older ones begin to falter.’[ix] The question of Digital Rights Management (DRM) haunts the field of ebooks. Demand (or patron) driven acquisition (PDA) of ebooks is one of the business model responses to this changing environment.

5.1 Pressure points for libraries

Patron or demand driven acquisitions was originally ‘conceived as a way of providing library users with a more time-efficient service than interlibrary loans (ILL). That idea came from librarians.’[x] Libraries are caught between two classic economic drivers. On the demand side the expectations from users are increasingly that the resources they require will be discoverable online (typically by search engines such as Google) and delivered immediately-free at the point of use. On the supply side libraries face diminishing resources. Libraries also find themselves in an increasingly competitive environment with commercial and non commercial 'library' services that provide alternative routes to resources, some of them free to use such as Google and Wikipedia. In the focus group a student commented:[xi]

‘Ever since I got a Kindle it’s a huge help... I love the library but it’s not good for all current student books – on Amazon it’s great, you get the option to buy books.’
In November 2011 Amazon launched its ebook loan service in the US which though limited in scope and focussed on ‘trade’ rather than scholarly publications, may well expand to provide some competition to academic library ebook loans. A focus group student said:

‘The books I want are never there [the library] so I sit at home and use Googlescholar/Google-books as they are there and free for general essays and also reading round subjects.’

These pressures motivate libraries to look at new business solutions. Writing in ‘Inside Higher Ed’ in October 2011, Steve Kolwich cited a recent (US) Education Advisory Board Report:

‘[The] report on the future of academic libraries identifies such demand-based services as an inevitable trend for libraries under pressure to prove that their expenditures are in line with their value. And one university says its own experimentation has produced damning data exposing the inefficiency of tradition collection-building compared to new methods that could prevail in the digital era. The Advisory Board**[xii]** report, a thick primer covering a range of trends in digital librarianship, predicts a shift in the way academic libraries provide book content to their patrons that mirrors a broader trend in digital media. (The report is not public.) Academic libraries will jettison “large collections of physical books in open stacks with low circulation,” the report says, in favor of licensing agreements with e-book vendors that will enable libraries to purchase only those books that are highest in demand, while paying short-term access fees for books that students use a little and nothing at all for books they do not use’**[xiii]**

In addition, with journals taking an increasing share of library budgets, the pressure is on to find ways to do deliver more effective ebook selection.

5.2 What is PDA?

PDA 'is a model of purchasing in which the librarians set the parameters of purchase and patron pull the trigger.' Whilst it comes in various forms, the essential characteristics or components of PDA environment are:-
  • The collection
The library provides access to user to a range of material that is outside a conventional notion of a library 'collection' in the sense that material is included that the library has not (yet) purchased or licensed. Michael Levine-Clark at the University of Denver believes that PAD (or DDA -Demand Driven Acquisitions as he prefers to call it) 'will force us to reconsider how we define the library collection'. He goes on to define the collection as 'the pool of titles available for potential lease or purchase-- and its size is bounded only by the library's budget.'[xiv] This 'consideration' pool of titles need active management. Whilst it is relative easy to fill initially 'a much more complicated aspect is the long term maintenance of title available....Libraries will need to craft rules for how long particular types of material remain in the pool.'[xv] This is certainly an issue that came out strongly our 'stakeholder interviews' and 'case studies' and metadata has a key role to play.

  • The user
Whilst users are rarely aware of being in a specific PDA context, user behaviour is linked more directly to acquisitions than it has been conventionally. This is often seen as helping make more effective purchases. Some evidence shows PDA acquired ebooks having more subsequent use than library selected ebooks. Some of the case studies and stakeholder interviews for this project supported this view. For example at Huddersfield:

Statistics of PDA show that student purchases are some of highest download books last year – higher than some of the library purchased items. A higher percentage of PDA e-book on PDA are used then non-PDA e-book, there is above average usage and also a smaller tail. 50% of e-books bought through Dawson were not used last year (average) – but only 40% purchased by PDA not used again.’**[xvi]** Letting the users select also has benefits for the library. The University of Huddersfield Case study asserted: ‘PDA reduces staff time spent on purchasing and selecting books’.

In, what might be termed, the 'strong' form of PDA, user behaviour (browsing a book for example) directly and automatically triggers the acquisition of the ebook. In its weaker form a library may mediate the PDA process. Elsevier’s ‘Evidence based’ model is not understood as PDA but it is a kind of patron driven way (based on usage data) to preselect an ebook collection that will should be better tailored to users’ needs.[xvii]

The nature of digital content makes it easier to disaggregate types of use. The time spent browsing, what chapters are accessed for example is much easier to record than in a print environment. This more granular approach to use opens up commensurate opportunities to make differential charges based. For example it creates opportunity to more easily employ a 'freemium' model whereby some use is free (e.g. short term browsing) and a scale of charges is offered for other kinds of use (e.g. short term loan, outright purchase).

  • The library
The library sets the parameters by which PDA works. The criteria that must be satisfied to trigger a purchase vary and a range of free and chargeable uses (e.g. browsing and ‘short-term-loan) may occur before a ‘purchase’ is made. According to Robert Johnson this is one of the most important aspects of the contract with the ebook provider: 'Remember, we're trying to delay purchase until we know patrons actually want this material'[xviii]
Although the material presented to the user is wider than material owned/licensed by the library it is nevertheless constrained by the library. ‘Libraries and e-book suppliers work together to identify titles that may be of interest.’ This limits what is available to be discovered.

6. Drivers: User motivations

One of the aims of the project brief is to help understand the motivation for patrons recommending or requesting the purchase of ebooks. In general patrons will not be aware they are initiating a PDA. Exceptionally, patrons may be aware they are engaged in PDA in those institutions such as the University of Newcastle[xix] that have initiated a mediated library approach in order to ration demand. PDA is currently used largely to help meet undergraduate needs. However Elsevier in their 'stakeholder interview' made the point about how ebooks have a role in research. Research undertaken on their behalf [xx] indicated that researchers ‘say they perform research outside their own functional areas to broaden the perspective’ and in terms of ‘fundamental knowledge’ books are more highly valued than journals. Our focus group included a number of postgraduate students who confirmed they used ebooks to a certain extent.

From the user's point of view 'selection' is simply differing forms of access to the content (such as browsing, downloading content (such as a chapters). These different forms of 'access' are used to contribute to a trigger point where the library buys the ebook. So at the heart of the various PDA models is a business model that takes a granular approach to usage and make differential charges. In this sense it is a version of the well understood 'freemium' business model whereby users (and really here the user is the library) gets a free basic service from the ebook platform provider that encompasses, for example, limited free browsing. The library gets charged for 'added' services. This may differ depending of the PDA platform provider. One approach is a 'short term loan' for which the library pays a (relatively small- percentage of book purchase price) fee. The steps (manifested by user's ‘clicks’) progress to the 'premium' service of full (or at least less restrictive--subject to the inevitable DRM) use initiated by a library purchase of (or actually a license for) the ebook

The motivations behind the use of ebooks were explored in a report commissioned by Elsevier.[xxi] This was a global study that looked at both user and librarian perceptions around the ebook format. For users the top positive perception was 'accessibility' followed by 'user-friendly', 'fast 'and 'convenient'. The project workshops with various stakeholders including librarians, ebook platform vendors and publishers confirmed this. Access, in particular off campus and 24/7 was seen as key. In essence the workshop stakeholders felt that the motivation for the end user is to simply get access to the best content for the job to hand. One librarian puts it this way[xxii]:

‘PDA expands the choice of what is available and makes material available instantly when the user wants it even if it’s out of hours. Many of our students live in rural areas and a substantial number of them are mature students who find it difficult to come to the library to access print material needed for their study. Students can access e-books when it suits them and can read around the subject through PDA’.

The University of Huddersfield Case study pointed out that:

‘The library provides e-books to address the common complaint from undergraduates that there are not enough books, e-books help as they can ensure at least one copy is available 24/7.’

The study remarked: ‘A benefit of PDA is its popularity with students, which may help with National Student Satisfaction surveys.

Electronic texts are inherently more flexible than print in terms of availability and flexibility (e.g. off campus and 24/7) A humanities student in the focus group commented:

‘I need to use a lot of books and so need to use lots of resources. It’s easier to keep research going if resources are electronic, it allows me to jump from one book to another’

Technically at least, a digital resource is 'non rival'. It can be used can be used simultaneously. The fact that one student is using an ebook does not necessarily stop another from using it. This attribute is of course one of the key concerns of publishers so constraints are added through forms of digital rights management (DRM): simultaneous use usually comes at a price.

6.1 The ‘problems’ or tasks users are trying to solve or address in their use of ebooks

Another way of looking at user motivation is to investigate the problems that a user is trying to solve. Or to put it another way the ‘jobs-to-be done’. The project addressed this issue using three different approaches. In the absence of resources to fund a large scale user study the rationale was to get to a consensus by a triangulation of different approaches.

The jobs-to-be-done (JTBD) approach is a well known methodology that helps organisations ‘get inside the heads’ of users or potential users to look at what real life tasks they are trying to do rather than a more abstract look at their ‘needs’. A workshop[xxiii] was help at the University of York with diverse group of participants -- librarians, ebook platform providers and publishers. The workshop identified the jobs as essentially:
  • Completion of a particular assignment
  • Preparation for an upcoming exam. This may include some reading around the subject.
  • Acquiring fundamental knowledge as part of the research process

The Elsevier study noted above cites this as a key job-to-be-done by researchers when they are outside their own functional areas. This analysis from the workshops was supported by the online survey the project instigated in cooperation with Dawson Books.[xxiv]




What task is this ebook being used for?
(Second survey results)
Answer Options
Response Percent
Response Count
Essay/assignment
47.1%
73
Research project
15.5%
24
Reading for a seminar
10.3%
16
Background reading
18.1%
28
Exam revision
1.9%
3
Other
7.1%
11
answered question
155
skipped question
1

Student interviews also confirmed this analysis. One student summarised the problems they were trying to solve by finding an ebooks as:
  • Essay/assignment
  • Background reading and improving own knowledge of subject.
  • Course reading generally

The focus group supported this view with some postgraduate students confirming too the use of ebook in their research. An undergraduate remarked that said they only used ebooks in their final year when researching for her dissertation.

6.2 Fundamental jobs

The JTBD methodology focuses attention on why the student is doing these tasks – what is the ‘fundamental’ job? The above tasks are clearly linked to the user’s overall academic motivation. For example does the student simply want (or to put it another way is the student prepared to devote the time and effort) to achieve a pass or do they want to attain the best possible result leading to a degree? The jobs above can therefore be subsumed into the ‘fundamental’ job of academic and research success. Students want to succeed (though they may have differing views about what success is) in their course and to do this they need to succeed in a number of different tasks such as assignments and exams. The Newcastle Case study said:

It’s still somewhat subjective but there is a sense that users have a greater choice of material with which to read around the subject. This in turn could lead to a better dissertation for example’.

Of course the university too wants to ensure the success of their students if they are to maintain retention rates which is linked to the funding they receive. Equally research funding received by an institution is linked to past research success. The evidence suggests that PDA is an effective selection tool in terms of re-use of resources.

Work lead by the University of Huddersfield also suggests a link between the use of resources (including e-resources) and final academic success. It is clearly in the students and university’s interest to get students to use more resources such as ebooks.[xxv]

6.3 Users

The workshop attendees were of the view that the main audience for library provided ebooks is undergraduates. Although a wide range of e-textbooks is still not widely available, most library ebook content is very closely aligned to courses and is consequently targeted at undergraduates. Nick Woolley from Kings College[xxvi] said the: ‘Present priority is support for taught courses -to improve deliver and range of ebooks’. As noted above the Elsevier study also identified researchers as users and the project end-user focus group provided some evidence of researcher use of ebooks However the workshop attendees felt they were very much a minority audience at present. In the online survey of those that said they were using the ebook for (postgraduate) research (just over 7% of the) PhD research barely registered.

Some libraries are exploiting PDA regimes to find out more about their users.

We also saw we might be able to get more detailed information on use. We had been following the PDA experiments in the states. We were already using EBL. EBL said they’d set up a patron information tool—so we could for example see which schools were using ebooks. When users authenticate for the first time for an ebook there are asked for relevant details (e.g. school)’**[xxvii]**

At the University of Wales Bangor
‘This year we are collecting information about the users. The first time they try to access any EBL e-book, users are asked for their school (e.g. social science) and whether they are undergraduate, postgraduate or staff. This information is associated with the patron’s Shibboleth ID so we can track what group of users are using e-books.’

6.4 Circumstance

The workshop attendees emphasised that users want relevant material ‘here and now’ with as few barriers placed in their way as possible. They want access 24/7 without being constrained by their location (e.g. off campus) or device. ‘Users want right format for [the] circumstance and context. E.g. [when] working from home or at a PC —[the] ebook may be best’. For some institutions number of ‘off-campus’ students is significant and giving those users easy access to resources is critical. By 2015, at the University of Liverpool for example: ‘ 15,000 students will be distance/international ones—higher than the number of on campus students...so every print book we buy is effectively denied to our online user—so ebooks made sense.[xxviii]

6.5 The role of PDA

PDA has a clear role in overcoming the barriers to access. It enables libraries to offers a range of material wider than the ‘conventional’ (purchased or licensed) collection. The University of Hertfordshire case study put it this way:

‘PDA enables the library to offer a wider choice of books, giving access to titles that we have not yet purchased. Unlike ILL (which is no longer heavily used), 'PDA' ebooks are in the library catalogue so are immediately discoverable and deliverable for use’.

In a conventional situation a title may be ‘in stock’ but demand may exceed the supply of copies. Workshop attendees noted that there may be in excess of 300 students on a module and at certain peak times competition for a specific resource is high. Students taking part in the focus group commented several times on the inability of the library to meet their demand with conventional print books. PDA helps meets such demand either through features such a free browsing or short terms loan before the library has to purchase outright additional copies. The Bangor case study said:

‘Demand can be patchy and we might find 30 copies of a recommended book sitting on the shelf most of the time yet there aren’t enough copies to meet peak demand. PDA and e-books in general help us to meet that kind of demand better.’

7. The role of metadata

As part of the JTBD workshop people were asked to look at metadata as a ‘solution’ to getting the job done and what capabilities metadata had to help solve the users ‘problem’. In summary the following metadata capabilities were indentified in the workshop sessions:

7.1 Metadata ‘capabilities’

  • In contrast to full-text indexing, metadata enable precision in terms of search/discovery.
  • A key capability of metadata is the filtering down of what is available. For example publication date can be a vital filter.
    • A key potential use of metadata is to help evaluate what the book is about. However there was considerable doubt over the value of ‘conventional’ metadata in this regard. Titles were felt to be important despite their sometime being not highly descriptive of the content. It was felt students rarely know about Library of Congress Subject Headings (LCSH). The consensus was that the value of such metadata may be what they can do ‘behind the scenes, for example in enabling links (using LCSH for example) to other resources in order to find relevant alternatives. It was thought Ebook platforms could make use of such data without necessarily displaying terms to users
    • Abstracts, chapter headings and product information were considered more useful from a user perspective in determining what the book is about to see if it is relevant to the job in hand.
    • Unique identifiers are needed to help to discriminate different ebook versions that have different ‘rights’. There might be an alternative version without or with laxer DRM for example.
    • In some cases images are very important and so metadata about whether ebooks have images or not could prove helpful.
    • It was felt, especially by librarians, that users would be reluctant to take time and effort to add ‘social metadata’ such as tags, reviews recommendations and ratings. However users might be prepared to use such data. This paradox could be resolved to some extent by deriving data automatically. For example recommendations can be derived from usage data and clickstreams.
    • Metadata about the user could be important so that the service knows who student is and the course they are studying and can therefore deliver more relevant results.
    • Usage and activity metadata could be used to drive additional services such as recommendations/suggestions. (For example it forms the basis for Elsevier’s Evidenced Based model for the acquisition of ebooks.
    • Metadata has a key role in scoping the collection that is available to the end user-- librarians set the preferences behind the scenes based on metadata.

7.2 The issue of quality

At the JTBD workshops there was some concern over the quality of metadata but it was muted. Libraries do get the records they need although there remain issues around timeliness. Overall there was a feeling that the situation has improved.[xxix] In the stakeholder interview Palgrave Macmillan said: ‘We see the value of metadata and have invested a lot in process’. Another publisher remarked; ‘The MARC record has become an integral part of the ebook offering. There were issues of quality at first as because publishers didn’t really understand MARC’. The Bangor case study supports the trend toward improved metadata quality:

On the whole the quality of records is good. We were going to upgrade MARC records for purchased e-books with OCLC records but on evaluation decided this might not be necessary. Judging by the level of usage, our students seem to have little trouble finding e-books in the catalogue and the discovery service.’
There can be a variety of fingers in the ebook metadata pie. The supply chain from publisher to end user can be quite complex and will typically involve data ‘crosswalks’ to convert ONIX data for example to library acceptable MARC records.
While libraries appear to be doing ‘as little as possible’[xxx] in terms of editing records, ebook platform providers do see a role in enhancing metadata to improve the service they give to libraries . For example as well as getting records from publishers, Dawson books records ‘are sourced from the British Library, Library of Congress, Bibliographic Data Services, or direct from the Dawson cataloguing team.’
This does add complexity. One vendor remarked: ‘We are all doing the metadata stuff in silos...great waste going on. We don't all need to be fixing metadata.’ This inefficiency was also noted in the 2009 ‘Creating catalogues’ report by RIN. They may be some rationalisation of the process as one of the impacts of the new library centric ‘discovery services.’[xxxi] NISO has recently launched the ‘Open Discovery Initiative’ to develop standards and recommended practices for library discovery services based on indexed search. [xxxii] This may spur on efficiencies in the ingestion of ebook metadata into discovery platforms.

7.3 Discovery

All stakeholders involved with, and interviewed for, the project agreed on the importance of metadata for discovery. However there is a variety of routes to discovery and users may encounter different metadata on their reading list, library catalogue, ebook platform, library discovery service or Google. The project’s online survey illuminated this diversity of approach.




How did you initially find out about this ebook?
(second survey results)
Answer Options
Response Percent
Response Count
Reading list
18.7%
29
Tutor recommendation
31.0%
48
Other recommendation - e.g. fellow student
2.6%
4
Library service (library catalogue, library discovery, databases)
31.0%
48
Ebook website (dawsonera)
11.6%
18
Google
0.6%
1
Amazon
0.6%
1
Other
3.9%
6
answered question
155
skipped question
1

It is interesting that, taken together, over half the respondents discovered about the ebook from reading lists, tutor and other recommendation rather than the library catalogue. This contradicts the views of some stakeholder we interviewed who saw the library catalogue as central to discovery. It may well be that the importance of the library catalogue is more as a ‘finding’ tool for known items ‘discovered’ elsewhere. A post graduate student at the focus group described a mixed approach in terms of resource discovery, finding and fulfilment:-:

‘I went onto library catalogue and the Metalib site to search databases and resources in this area. This came up with list of journals and books, I found one I liked –the title sounded perfect - then I clicked on it to find out if it was at York but wasn’t. I copied the details into Google and found it on free on Google books’.
The academics we interviewed were more certain of the importance of reading lists and tutor recommendations for discovery with one pointing out:

The reading list is a key tool to enable the discovery of the right material. It also provides annotations about which readings are most important—i.e. that week or in general. An academic might also indicate (verbally) preferred reading in lectures and suggest particular resources for a specific assignment. We looked at 3 modules last year and asked students what they used: material from reading lists or material they have found themselves. Most usage came from reading list.

There is clear opportunity here for the closer integration between reading lists, which are seen as largely in the domain of academics and library discovery and management systems. This had had a long history but has not resulted in very satisfactory solutions[xxxiii]. Libraries are conscious of this and are either building their own reading list systems (e.g. University Huddersfield) or implementing commercial systems such as Talis Aspire.

7.3.1 The impact of library centric discovery services

New library centric discovery services will undoubtedly have a major impact. They have achieved relatively rapid adoption in university libraries around the world.[xxxiv] It seems likely that the management of ebook metadata will move away from the local library systems and into the new cloud based ‘web scale’ library centric discovery services. This will make ebook metadata analogous to journal article metadata in that the data will not be created, stored or managed within a library management system environment. Discovery services are already ingesting large quantities of ebook metadata. NISO has launched an ‘Open Discovery Initiative’ to improve the flows of metadata (for all resources –not just ebooks) from publishers into the huge centralised discovery service indexes. Taking part in the initiative would appear to be a great opportunity for JISC to influence developments. Writing about the initiative Marshall Breeding comments:

‘The fundamental premise of these products involves a transfer of content from publishers to discovery service producers. Rather than having many different ad hoc arrangements, there would be significant efficiencies gained in developing standards or best practices in how these data are transferred. For example, such a standard would save publishers from having to create different mechanisms for each discovery provider. Such standards would be especially beneficial to discovery providers since they may work with hundreds of content providers’.

It is remains unclear how libraries will delimit their collections available for PDA in this new context. In a December 2011 article in ‘Against the Grain’[xxxv] Peter McCracken made the following observation:

‘The rapid introduction of Web-scale discovery layers, such as Serials Solutions’ Summon,1 EBSCO’s EBSCO Discovery Service, Ex Libris’ Primo Central, OCLC’s WorldCat Local, and others, provides the perfect layer for applying an a la carte approach across large swaths of data. All of the pieces of technology are available to make this work; all that’s needed is some modifications to administrative interfaces in discovery layers, some additional data tracking and reporting, and most importantly, a willingness among libraries and content providers to try something new.

McCracken is talking about e-resources in general but his observations may well have particular resonance in the context of ebooks and PDA. PDA is in effect a variant on the ‘Pay-per-view’ model. He goes on to describe how Discovery services might be enhanced especially to track Pay-Per-View (PPV) and Pay-Per-Click (PPC) traffic.

‘Discovery layers must build administrative tools that allow them to track PPV and PPC statistics and fees for each database in their collection, track which databases are managed in what fashion by library, track discounts offered by content providers to libraries, bill libraries for usage on a monthly or quarterly basis and distribute funds to content providers on a similar schedule, and much more.’

7.4 Selection and Acquisition

One of the key problems for libraries in the adoption of PDA is that they run out of money too soon. For example at Bangor:

Last year we started in October and ran out of money we had set aside (£60k) by February. In order to control demand we suppressed the records from the catalogue for those e-books that we had not already purchased.’[xxxvi]

The Newcastle case study neatly describes the increasing demand and the incremental steps taken to control it to keep within budget:

‘We launched [PDA] quite quietly but demand rose quickly and we couldn’t sustain the high level of demand. We launched in February 2010 and saw a steep rise in acquisitions over the first 3 weeks. We put in a limit of 3 loans per day. From week six the loan limit per day was set to 1 and we altered the auto-purchase trigger from 3 loans to 5. We then started to display the cost of loans to user and we think it acted as a moderate brake. From week 18/20 we started to mediate all requests. From about week 20 we started to do some profiling to limit the resources that could be discovered’.

Libraries therefore use metadata to exert considerable control by profiling what records are made available in the first place or may delete records afterwards so they cannot be discovered. The University of Hertfordshire Case study reports:

‘we anticipate that a key role for metadata will be as a tool for the library to balance the competing demands of users for a wide range of material to discover and the need for the library to constrain demand within budgetary limits.’

Metadata used for profiling may include subject, publication date (i.e. not before yyyy), language, country of publication (e.g. ‘no US publications). The University of Hertfordshire is ‘looking into profiling the available collection by using Dewey range, date (to weed out 'old' material) and publisher/country of publication—(for example, for Law we generally don't want old editions or US titles).’ It appears that metadata will continue to play the key role in determining the ‘collection’ made available to end users in terms of PDA. However it is not without its issues.

7.4.1 Identifiers

Moving data in and out of library catalogues is not always straightforward. As well as issues of time spent processing the data through the library management system there are issues around ebook identifiers’

‘Working with MARC for ebooks is always difficult and PDA is no exception. The lack of consistent unique identifiers undermined any possibility of 100% success in matching records for ebooks from other sources to avoid duplicate acquisition. In this regard you have to gauge when to cut your losses.[xxxvii]

It becomes increasingly important as libraries buy a variety of ebook solutions from a number of different vendors. As one of the case studies pointed out:

‘Issues of record matching can be significant where one has different ebook collections. For example we bought a Springer ebook package (c 23k records) –paid for as a collection and not as a PDA resource. This has caused some issues with how to deal with duplicate records. There is not universal and unambiguous identifier in use as yet.

In 2008 Nielsen which, in the UK and some other geographies, has the role of the official ISBN agency restated the position of the International ISBN Agency on the use of ISBNs for ebooks.
In trying to motivate publishers to adopt this position Nielsen issued its own policy statement that ebooks will only be listed on the Nielsen Book database and in Nielsen Book products and services if they are identified in accordance with the ISBN standard. Publishers do not always adhere to the Nielsen policy.

The issue of identifiers is critical and will become more important as ebook content becomes fragmented, as many believe it will, into chapters for example. In this case a chapter will become a discrete ‘tradable product’ and will need its own identifier. The BIC (book Industry Communications) Metadata Futures Group was set up to address the future demands of the industry in terms of systems design and identification. It recognises that with the growth of digital distribution there has been a corresponding increase in the number of differential products, often superficially identical, which systems and other existing resources have had to cope with.[xxxviii]

7.5 Evaluation

A key activity within the selection process is evaluation. One of the ebook platform providers we spoke to describe the roles of metadata this way: ‘First roles are for search and discovery and second is evaluation --is this a book I want to spend time with?’ There is a growing body of evidence, supported in some of our case studies, which PDA leads to more effective selection. We might therefore assume that, in part this was because the user had successfully evaluated the ebook before purchase. Our online survey found that around 70 per cent of users felt the ebook had met their needs. Just over a quarter remained ‘unsure’.

In general the stakeholders we interviewed were equivocal about the role of metadata in evaluation. The University of Newcastle Case Study is typical:

‘We are not sure of the role of metadata displayed to users in evaluating the resource and therefore whether it is a factor that determines use.’

At the workshops one librarian pointed out that the library catalogue user at their university is presented initially with a brief record that contains only basic metadata. She felt that users went from that point to the ebook itself and made their evaluation there. Asked if they thought metadata was useful for evaluation an academic said: ‘I don’t think so—at least not what they see in a library catalogue. The reading list is more helpful from this point of view.’

The ebook platforms typically include chapter titles (TOC) and product descriptions and this is now increasingly included in the MARC record provided to libraries for loading into their catalogue. As we noted earlier this enriched metadata has value in terms of improved discovery. It also seems to have value when displayed to the end user. In one of our interviews a student was asked how they decide whether to use an ebook and what information they use to help their decision.

Chapter titles, publication dates newest makes a big difference in my subject area (science), summary details or abstract are useful to make a decision, I might find it useful to flick through a few pages BUT Table of Contents was the most useful section to make a decision’

In one of our librarian interviews the utility of metadata for DRM was considered: ‘helping people know how they can use resource. Click ‘what can I do with this?’ The JISC funded RELI project has looked at this issue but it was based on the assumed wide adoption of the ONIX-PL licence which has not made great progress. [xxxix]

We asked users in the online survey: ‘When choosing this ebook to use what information about the book was most useful to you in deciding to select it?’ As can be seen the top three elements were title, blurb and indication it was on a reading list.

When choosing this ebook to use what information about the book was most useful to you in deciding to select it? Please rank the following in order of priority with one being the most important
Answer Options
1
2
3
4
5
6
7
8
9
10
Rating Average
Response Count
Title
38
22
19
5
5
2
8
7
6
6
3.58
118
Indication that it was on a reading list
29
18
13
11
9
6
6
5
1
12
4.00
110
Author
4
15
17
15
15
16
10
11
5
1
4.85
109
Date of publication
1
12
23
19
8
17
7
7
4
6
4.93
104
Information on what the book is about (blurb, summary/abstract)
21
13
20
17
10
11
7
4
3
2
3.88
108
Table of contents
2
9
9
17
21
20
10
6
4
3
5.19
101
Subject keywords
11
9
11
10
21
16
24
5
3
2
4.98
112
Publisher
3
11
4
6
3
6
9
32
24
7
6.83
105
Price
11
7
2
4
10
4
6
13
37
15
6.88
109
Other
9
3
2
5
4
3
7
7
6
43
7.51
89
answered question
155
skipped question
2

7.5.1 Social metadata

Social metadata such as tagging, reviews, ratings and recommendations is a key evaluative component of many web based services but has only recently begun in library systems. A student we interviewed said: ‘People who read this also read.... would be a good feature to include’ The University of Huddersfield has been a pioneer in this regard and pointed out in their case study:
‘We have use circulation histories to build a recommender. We can track this right down to module level and can build recommendation lists (so on OPAC will have an entry at the bottom of the catalogue record saying ‘people who borrowed this also used xxxx’
ExLibris offer the bX recommender service for e-journal articles as part of the Primo discovery service and we can expect to see such features being made available from other vendors. The commercial discovery services are building huge central indexes of e-resources which are beginning to include ebooks. Vendors such as Serial Solutions/ProQuest and EBSCO own ebook platform services and are indexing that content already. There are major opportunities for such vendors to combine information about the resources themselves with usage data derived from clickstreams to deliver new services based on such ‘collective intelligence’. One of the academics we interviewed declared social media is ‘where the innovation will come’ and went on:
Tagging is so widespread and current in the wider world —so it’s what people expect. We already add notes to reading list–e.g. ‘Chapter 4 particularly good on gender’. This notion could be expanded to include tagging by others, including students’.
There are wider and more complex issues too involving the opportunities digital format have for sharing functions like annotations. An academic said:

‘I would like the market to make moves to providing a better user experience: the ability to use the content—make notes, annotations etc..without being tied to a specific ebook platform, or having to be online’.

Some doubts about the take up of social features were expressed by librarians at the workshop sessions. There was doubt that users would be prepared to put the effort in to add tags, ratings or reviews. This was supported to some extend in the end user focus group. However this may be dependent on how such features are ‘packaged’. One of the students who use their Kindle a lot remarked:

‘I try to tap into the readers view point and what they use the books for. I use the comments from other readers and I would like to see this in the University – I would find this a useful experience’.

Some of the initiatives in the ‘trade’ ebook sector might be adapted to transfer to the scholarly market. For example the Kobo reader ‘Pulse’ feature[xl] enables users to:

‘• Feel the life of a book: Watch a book’s pulse get stronger as comments are added and more readers are reading!
• Get more out of reading: See what others thought of what you just read and get interesting stats, such as how many others are reading with you.
• Connect with others: Connect with friends and a community of almost 5 million Kobo readers around the world!
• Join the conversation: Like or dislike a book, comment on memorable passages and contribute to ongoing discussions while you read.’

At the Frankfurt Bookfair in October 2011 NISO held workshops on ‘E-Book Annotation Sharing and Social Reading.’[xli] These workshops discussed issues surrounding bookmarking and annotation and possible areas for standards.

8. Conclusions and recommendation

8.1 The market

Patron/Demand driven acquisitions has only been adopted in UK Higher Education over the last 2-3 years. None of our case studies has more than two years experience and most have only one full year. The market is still dynamic and evolving towards achieving sustainable models. Other different approaches such as Elsevier’s Evidenced based Selection approach is also being evaluated by some UK university libraries. Liverpool University’s experience is included as one of the project case studies. JISC can support the better operation of the market by facilitating the exchange of information. The Case studies included in this project are a part of this information exchange.

8.2 Metadata

JISC should use its influence in bodies such as NISO, and in particular the Open Discovery initiative to help create the most effective and efficient environment for ebook. In addition there is opportunity for the JISC to coordinate efforts with organisations such as UKSG (a body encompassing all stakeholders) to clarify the mechanics of how ebook metadata might best be managed and the enhancement needed by discovery services to enable the best implementation of business models such as PDA.

8.2.1 Identifiers

The growing importance of identifier s for ebook is recognised within the industry. We have already mentioned the BIC metadata futures group//**[xlii]**// and JISC should engage with this group to achieve the best possible outcome for UK HE

8.2.2 Social metadata

One of the academics we interviewed said that ‘Social media is where the innovation will come‘ and this project highlights opportunity for further work to investigate the impact and opportunities of social metadata. As we pointed out JISC has already funded relevant projects and there are a number of existing initiatives which JISC can contribute to. There are some complexities: ‘social media and metadata cover a number of different things including rating, reviews, tagging annotations etc. Some may be explicitly generated by users (e.g. reviews) whilst others may be generated automatically (e.g. recommender services based on past use). Importantly too it includes metadata about the user either explicitly expressed (for example in terms of their course) or implicit in their online behaviour expressed, for example, as ‘clickstreams’.










References

The project wiki contains further detail including case studies and other material that can be freely downloaded and reused:-

Case Studies
https://ebmotmet.wikispaces.com/Case_studies
Stakeholder interviews
https://ebmotmet.wikispaces.com/Stakeholder_interviews
Workshops
https://ebmotmet.wikispaces.com/JTBD_workshops
User survey
https://ebmotmet.wikispaces.com/User+Survey
End user focus group
https://ebmotmet.wikispaces.com/Focus_group
Other relevant work
https://ebmotmet.wikispaces.com/Synthesis


[i] Website http://observatory.jiscebooks.org/
JISC national e-book observatory project. Key findings & recommendations: JISC Collections. November 2009
http://observatory.jiscebooks.org/files/2011/08/JISC-national-ebooks-observatory-project-final-report.pdf

[ii] http://fe.jiscebooks.org/ ‘The e-books for FE Project will provide Further Education colleges in the UK with access to a Core Collection of e-books on a platform which provides functionality suitable for the needs of the community. The e-books are available under the terms of the JISC model licence e.g. Staff and students can incorporate parts in teaching learning materials whether electronic or paper. This includes course packs, presentations, VLEs, mobile devices, project work and much more.’

[iii] More details concerning the project team are available of the project wiki ‘project team’ page https://ebmotmet.wikispaces.com/Project_Team

[iv] https://ebmotmet.wikispaces.com/

[v] For details see http://touchpress.com/titles/thewasteland/

[vi] Patron-Driven Acquisition of publisher hosted-hosted content: bypassing DRM.' By Jason Price. Against the Grain June 2011. Volume 23. Number 3 ISSN 1043-2094

[vii] The 2011 Horizon Report. The New Media Consortium ISBN 978-0-9828290-5-9
http://net.educause.edu/ir/library/pdf/HR2011.pdf

[viii] ‘Goods that are non-rival are goods that can be enjoyed simultaneously by an unlimited number of consumers’. Wikipedia 'Rivalry (economics) http://en.wikipedia.org/wiki/Rivalry_(economics)

[ix] The 2011 Horizon Report. The New Media Consortium ISBN 978-0-9828290-5-9
http://net.educause.edu/ir/library/pdf/HR2011.pdf

[x] E-books procurement a disruptive business. By Elspeth Hyams. CILIP update with Gazette. April 2011 (interview with Kari Paulson of ebook aggregator EBL)

[xi] The focus group report: https://ebmotmet.wikispaces.com/Focus_group

[xii] Education Advisory Board http://www.educationadvisoryboard.com/

[xiii] 'P.D.A. in the Library'. By Steve Kolowich. Inside Higher Ed 28th October 2011
http://www.insidehighered.com/news/2011/10/28/e-book-acquisition-based-use-and-demand-could-save-libraries-thousands

[xiv] Developing a model for long-term management of demand driven acquisitions.' By Michael Levine-Clark. Against the Grain June 2011. Volume 23. Number 3 ISSN 1043-2094

[xv] Developing a model for long-term management of demand driven acquisitions.' By Michael Levine-Clark. Against the Grain June 2011. Volume 23. Number 3 ISSN 1043-2094

[xvi] University of Huddersfield Case Study

[xvii] The project case study from Liverpool was based on the Elsevier Evidenced based model

[xviii] Purchasing options in Patron-Driven-Acquisitions'. By Robert Johnson. Against the Grain June 2011. Volume 23. Number 3 ISSN 1043-2094

[xix] See the University of Newcastle Case Study that is an appendix to this report

[xx] '2009 Global Online Books Study: Librarians and Researchers.'' Satisfaction Management systems (SMS) for Elsevier. Final report 31st March 2009

[xxi] '2009 Global Online Books Study: Librarians and Researchers.'' Satisfaction Management systems (SMS) for Elsevier. Final report 31st March 2009

[xxii] See the Bangor Case Study

[xxiii] Specific details of the ‘Jobs-to-be-done’ approach and the workshops is on the project wiki at https://ebmotmet.wikispaces.com/JTBD_workshops

[xxiv] The survey was done in two ‘tranches’ over October and November 2011. The second (larger) tranche had an additional question to clarify what was meant by using the book for ‘Research’. Other results for both tranches were very similar

[xxv] The Library Impact Data Project (February to July 2011), The JISC project website
http://www.jisc.ac.uk/whatwedo/programmes/inf11/activitydata/libraryimpact.aspx
Presentation of findings http://eprints.hud.ac.uk/11659/

[xxvi] Librarian stakeholder interview

[xxvii] University of Newcastle Case Study

[xxviii] Librarian Stakeholder Interview with Martin Wolf of Liverpool University

[xxix] ‘Vision for E-Book Metadata. By Hazel Woodward and Paul Needham (2010?) JISC Resource Discovery task force (RDTF) document (?).In this paper recommended actions for JISC included: ‘Ensure that in all negotiations with commercial vendors (publishers and aggregators) the requirement for high quality metadata is an integral part of the agreement’

http://www.docstoc.com/docs/76037904/Hazel-Woodward---Vision-for-E-Book-Metadata

[xxx] University of Durham, Librarian Stakeholder interview

[xxxi] For an overview of library Discovery services see the article on SCONUL’s Higher Education Library Technology (HELibTech) website (wiki) http://helibtech.com/Discovery

[xxxii] ‘NISO Launches New Open Discovery Initiative to Develop Standards and Recommended Practices for Library Discovery Services Based on Indexed Search’. NISO Press Release 25 October 2011
http://www.niso.org/news/pr/view?item_key=21d5364c586575fd5d4dd408f17c5dc062b1ef5f

[xxxiii] See ‘A perspective on resource list management.‘ Ken Chad. Library & Information Update (p.39-41). CILIP June 2010.
http://www.kenchadconsulting.com/wp-content/uploads/2010/07/Perspective_on_Resourcelist_Manageemnt_CILIPUpdate_June2010.pdf

[xxxiv] For general information about discovery services and the UK situation see SCONUL’s Higher Education Library Technology (HELibTech) wiki’s ‘Discovery’ page http://helibtech.com/Discovery

[xxxv] ‘A Proposal to Improve and Expand Access to Electronic Resources through Per-Use Pricing’. By Peter McCracken. Against the Grain v23 #5. December 2011

[xxxvi] Bangor Case study

[xxxvii] Kings College Case study

[xxxviii] The BIC Metadata Futures Group http://www.bic.org.uk/77/METADATA-FUTURES-GROUP/ The June 2011 Discussion Paper by BIS’s Francis Cave set out some of the complexities involved http://www.bic.org.uk/files/pdfs/110617%20Discussion%20paper%20-%20metadata%20description%20of%20assets%20products%20and%20groupings.pdf

[xxxix] See ‘RELI a project to pilot the development of a licence registry: final report’. (2009?). The RELI project looked at a demonstrator for ‘a Licence Registry service that would interpret library licences for libraries encoded in ONIX-PL in such a way that a user, typically a librarian or an end user, could see at a glance what permissions are granted to the users for, e.g., downloading or redisseminating content.’
http://www.jisc.ac.uk/whatwedo/programmes/reppres/sharedservices/licence.aspx

[xl] ‘Kobo brings books to life with kobo pulse’ by Anna Bittner. Kobo Blog. 27th September 2011. http://blog.kobobooks.com/kobo-brings-books-to-life-with-kobo-pulse%E2%84%A2/

[xli] ‘E-Book Annotation Sharing and Social Reading’. NISO workshops. October 2011 http://www.niso.org/topics/ccm/e-book_annotation/

[xlii] http://www.bic.org.uk/77/Metadata-Futures-Group/