Books Books
Image via Wikipedia

The gods of serendipity blessed me recently and brought to my attention an outstanding document prepared by members of the staff of the Cornell University Library.  The committee was chaired by Kizer Walker; I am thus calling the document the Walker Report, though that phrase never appears in the study.  This is an internal document, but librarians being librarians, it is posted on a publicly accessible Web site.

Yes, I got permission from Walker to post the link.

The document summarizes a study of book circulation at Cornell.  It is carefully done, data-rich, and full of surprises.  It also is inadvertently amusing — knowing that the data could easily be misinterpreted, it manages to slip in a cautionary note (“Don’t jump to conclusions”) every couple paragraphs.  This is weird; one would hope (vainly, apparently) that in a university people would feel confident about expressing their views and meeting with intelligent reception.  Heh.  In any event, anyone interested in scholarly communications should take a look at this.  It would be great if similar studies were conducted elsewhere, and better yet if the data were aggregated and mined for patterns.

The aim of the Walker Report is to get a better handle on how books actually circulate in order to plan collections better.  Are there too many books?  Too few?  Perhaps there are too many in some disciplines, not enough in others?  How could this information be used to build plans for migrating toward patron-driven acquisition (PDA), which potentially could lower libraries’ costs — or, better yet, align collections with usage such that library funds are most efficiently spent?  I speculated about PDA a couple years ago, though I had not yet stumbled upon the term (I called it “purchasing on demand”).  My interest was in how PDA would affect the overall value chain.  PDA might save money for librarians, but could create headaches for library vendors (e.g., Yankee Book Peddler) and the publishers themselves, some of whom would not survive yet another reduction in unit sales for monographs.  Publishers are going to want to look very closely at the Walker Report, as it is beginning a process that will ultimately reform their industry.

Before I go any further, I want to be on record that I don’t view the circulation of scholarly materials as a popularity contest.  It’s wonderful that some books circulate often, but a book that circulates rarely, and even books that circulate not at all, have value.  That value is not necessarily captured in commercial terms. A book on the shelf can sit there just in case, and just in case is a good reason.  And the very writing of a book creates value independently of its reception, as the writing process inevitably bring others into the author’s orbit and trains a professional.  The ultimate measure of a book is not where it stands on the bestseller list, but where it stands in the web of relationships among scholars, their profession, and their institutions.

The nagging question, however, is that at a time when every expenditure is being scrutinized so closely, will simple circulation counts come to drive not only specific purchases but libraries’ overall materials budgets?

Some highlights from the report — but, please, the real merits of the study are in the details.  For all books published since 1990 that are now in the Cornell collection, 55% have never circulated one time.  Is that a big number or a small number?  It depends.  But Cornell has collections in many languages; for English, the number drops to 39%.  Presumably given more time (that is, more than the 20-year period of the study), the rate of circulation would rise, but another interesting item is that circulation tends to level off after 12 years.  This prompts me to ask for a show of hands:  How many publishers in the room prepare sales forecasts before acquiring a book with a 12-year horizon in mind?  With PDA, you will have to.

An interesting aspect of the analysis is the “snapshot” methodology, in which circulation records are examined for all the books that are out on any single day. Librarians will be able to make more of the snapshots than I can, but I was intrigued to learn that when a snapshot was taken, the average undergraduate each had one book in circulation.

Of course, circulation varies by discipline (philosophy gets very good grades) and presumably by publisher and author as well, though those last two fields were not captured in the report.  But they will be in future iterations.  Publishers should consider this:  the bright light of computer-assisted quantitative analysis is being brought to bear on their brands.  Which brands will benefit from this? Which will suffer? And how many book publishers budget money for brand marketing as journals publishers do?

Scholarly book publishing is becoming more like journals publishing every day.  Books are going digital; journals are already there.  Books are now being sold in aggregations; journals have been sold that way for some time.  Books are going to be sold by subscription; subscription is the prevailing form of marketing for journals.  And books are going to be subject to various metrics (pageviews, downloads, whatever) to determine whether they belong in a library’s collection; journals have been living with that situation for some time.  Indeed, one way to look at the growth of interest by publishers in digital book aggregations is as a way of fending off the implications of PDA and quantitative measurements of publishers’ brands — a weaker title can hide in a collection of better titles, just as second-tier journals retain their circulation because they are protected by the sheltering embrace of a Big Deal.

The next battle for scholarly book publishers thus is not in getting their books into digital form — that’s a given; that’s the last war.  The next battle is brand-building after the sale has been made.  This is aftermarketing, and it’s going to require new tactics and resources.  Publisher X is going to be thinking about driving up usage and circulation statistics for books that are already in library collections.  Good stats will increase the likelihood that a library will purchase the publisher’s next book, and the next after that, and so on.  Simply by exposing the data, beginning at Cornell with the Walker Report, like the butterfly that flaps its wings on the other side of the world, a series of changes is initiated leading to the emergence of the importance of publishers’ brands.  Yes, it’s a lot easier to play this game if you start with a strong brand to begin with.

Enhanced by Zemanta
Joseph Esposito

Joseph Esposito

Joe Esposito is a management consultant for the publishing and digital services industries. Joe focuses on organizational strategy and new business development. He is active in both the for-profit and not-for-profit areas.

Discussion

16 Thoughts on "Aftermarketing: What Publishers Now Must Do After the Sale"

I think that in certain other industries, the term “post-marketing surveillance” is also used to describe “aftermarketing,” particularly with pharmaceuticals. The latter has an espionage overtone, unfortunately, but also underscores data collection and product utility after purchase. These also sound a little grim….

I think you’re misinterpreting the data in your statement that “the average undergraduate each had one book in circulation” on one of the snapshot days. (What is an “average undergraduate,” anyways?) It would be more accurate to say that, on average, one book was checked out for each undergraduate.

I have only read the Executive Summary so far, but will want to read this study in full soon, as it appears to be a model of its kind. But I do have some preliminary thoughts in response to Joe’s own presentation.

I am glad that the data are broken down into categories of user, and it should be no surprise that monograph circulation is higher among graduate students and faculty than among undergraduates. If “monograph” means here what university presses generally publish under that rubric (and I didn’t see any definition of the term in the report on first glance), then it makes perfect sense that undergraduate use would not be high because monographs are high-level scholarship not aimed at undergraduates at all. Still, one can’t fully evaluate what these data mean unless one knows what the range of “monograph” covers. I suppose the study doesn’t include any fiction, poetry, etc. at all. It might be interesting to know what circulation for those types of books is, for comparison’s sake.

And what is meant by “usage” anyway? Circulation count is a very blunt instrument to measure the value of a collection, as Joe suggests. One would like to know just what these “monographs” were used for, by the different categories of users. And does the study make any effort to include in-library usage, which surely is important in evaluating whether any given book has actually had some intellectual impact in the university?

Joe rightly points to the importance of branding, but this can have perverse effects, too, as it has in the P&T process where the “prestige” of a publisher’s imprint can sometimes be given too much weight and where presses in business for hundreds of years (Cambridge and Oxford) have an unfair advantage over presses that may be cultivating new fields or doing very good work in old fields but not for long enough to have acquired the “prestige” of the older presses.

There is one way for the newer and smaller presses to level the playing field, however, and that is by winning prizes from the major scholarly associations. There is no better way, in my opinion, for a smaller press to enhance its reputation in the marketplace, and with the advent of PDA, this may be even more important in the future, as PDA orders may be stimulated by the announcement of major book awards. I made a concerted effort while director at Penn State to win book prizes for just this reason. In its entire 33-year history before I became director, the press had won just four prizes. By the time I left 20 years later, the tally was well over one hundred. That is “aftermarketing.”

As the principal data analyst for the Walker Report, I can address some of the issues you raise. “Monograph” as used in our report, is defined quite broadly. Anything with a MARC Bibliographic Level of “monograph/item” and a MARC Type of Record of “language material” was included. The designation excludes serials, but it absolutely does include fiction, poetry, textbooks, and plenty of other print material of potential interest to undergraduates (to the extent that undergraduates are interested in print at all).

We do not assert that circulation alone measures the value of a collection, and we do acknowledge that there are other forms of usage, including “in-library” use. Pragmatically, however, circulation is the most reliable surrogate for usage available. While we have in-library usage data, it’s an even blunter instrument than circulation, lacking both consistency and specificity.

Meanwhile, considerable effort was made to refine the circulation data. For example, all non-patron circulation, such as routine placement on new book shelves, or use by a digitization project, was filtered out. In examining circulation patterns for the broader collection, all monographs in non-circulating locations were excluded.

It would be interesting to know how the books were used by different categories of users. Circulation data can’t distinguish leisure reading from research use, nor fact-checking from deep scholarship (though a course reserve book that circulates for two hours at a time is probably not being used to write a thesis). We do have some data that adds shades of nuance to the circulation picture, such as snapshot results that cross-reference users’ departmental and field of study affiliations with the characteristics of the books they check out. Ultimately, however, any large-scale study of automatically generated transaction data will be limited to answering questions of what, when, and where (and to a lesser extent, who). We’ll leave it to ethnographic researchers to uncover the hows and whys of library collection use.

You say, “For all books published since 1990 that are now in the Cornell collection, 55% have never circulated one time. Is that a big number or a small number? It depends. But Cornell has collections in many languages; for English, the number drops to 39%.”

The report actually states that 61% of the English language books had circulated since 1990 (see p. 2 and p. 14 of the report).

These are the same numbers. You are citing how many books circulated, I am citing how many never circulated. The two figures total 100%.

Joe, you suggest that publishers are going to have to start thinking about how to drive up “usage and circulation statistics for books that are already in library collections.” I am sure that most librarians would be thrilled if usage and circulation of the materials they bought went up. Do you have any suggestions on how this might happen?

There is no single answer to this question–and I should add that this is the kind of question that consultants live on. To come up with an answer, you would have to build a matrix that would include such factors as medium (print vs. electronic), subject area, price range, etc. A strategy for OUP will differ from one for Penn State Press, which in turn will differ from Random House’s.

That said, there are some general points to bear in mind, the most important of which is a complete rethinking of a publisher’s metadata strategy. Marketing is based on metadata, but book publishers don’t uniformly do a good job in creating and maintaining it. And the metadata has to be consumer metadata, since it is individuals, not information professionals, that have to be reached.

Since ultimately the publisher is the best source of metadata, we will begin to see more publishers attempting to populate library systems with metadata directly. This in turn will lead more publishers to try to sell directly to libraries, especially for ebooks.

An interesting question is the intersection of libraries’ and publishers’ interests on the discovery matter. There is room for thoughtful collaboration in this area.

Getting book prizes is one way of boosting usage and circulation, since prizes are often awarded to books a year or more after they have been published, by which time many libraries will already have them in their collections (at least under the old approval-plan system). Reviews in high-profile media are another way. Neither prizes nor reviews can be manufactured by publishers, of course, but they are unlikely to happen with significant publisher effort either. I’m not sure how prizes and reviews can be entered into metadata, however.

Publishers adept at the use of ONIX can disseminate information about prizes throughout the value chain. Of course, not all recipients of ONIX metadata are equipped to use it.

Whether reviews affect library use is one of those vexing questions that is very difficult to answer. Some people at Colorado have been doing good work on the issue. See, for example, Jobe, Margaret M. and Michael Levine-Clark, “Use and Non-Use of Choice-Reviewed Titles in Undergraduate Libraries,” Journal of Academic Librarianship 34, no. 4 (2008): 295-304, and Levine-Clark, Michael and Margaret M. Jobe, “Do Reviews Matter? An Analysis of Usage and Holdings of Choice-Reviewed Titles within a Consortium,” Journal of Academic Librarianship 33, no. 6 (2007): 639-646.

Joe is right, though, that it wouldn’t hurt if publishers got review citations added to the records in WorldCat Local.

I wasn’t thinking of Choice reviews specifically, though of course Choice’s publisher assures publishers that Choice reviews are used by librarians to plug gaps in their collections. I was thinking more of high-profile reviews in media like the NYTBR, NYRB, TLS, TNR, Economist, Atlantic, etc. that would be more likely to drive PDA requests.

Awareness is key, and anything the publisher can do to help the potential reader realize that the material exists will help usage numbers. Getting the material into regularly searched databases like Medline, and having it indexed and made available (at least in small snippets) by Google are things a publisher can do to help guide readers to books.

Adroit use of Google AdWords can also be successful, especially if the subject of the book is such as to be singled out by use of a term that is fairly targeted to the book’s content and not likely to be used by many others, thus keeping the ad price reasonable. We did this at Penn State Press, too.

As the CHOICE publisher mentioned in one of the comments above, I would add simply that the studies I have seen suggest two things about the impact of CHOICE reviews in academic library settings: 1) on average, and all other things being equal, a CHOICE review means academic libraries will purchase more copies of the title in question, and 2) titles reviewed by CHOICE are less likely to never circulate than other titles, although their average circulation rate may not be noticeably higher. In any case, one option that a goodly number of academic libraries have exercised in recent years is to enhance their OPACS (once known as “Amazoning the OPAC”) with metadata, often including CHOICE reviews, obtained from either Content Cafe (a Baker & Taylor product) or Syndetic Solutions (a ProQuest/Bowker service). The objective is clearly to boost circulation by providing the OPAC user with additional, useful, information about the title. While I have not myself seen a study that evaluates the effectiveness of this effort, it seems to me that publishers for whom academic libraries are an important market, might consider providing a similar service for their library customers. And if anyone has seen a study of the impact of OPAC enrichment on monograph circulation, I’d love to know about it.

Comments are closed.