The JSTOR White Paper on Reimagining the monograph quotes a report on print collection use at Cornell libraries. They found that by 2010 55% of the books published since 1990 had not circulated. In the years 1990 to 2010 Cornell’s library acquired 1,654,034 print monographs, 55% of them in English (not, one assumes, the same 55%).

The report found that “Most books in circulation on April 19, 2010, were charged to graduate students, who accounted for 34% of the total charges. Faculty had out another 23.6%. Undergraduates had out only 10.7% of the books charged – 16,744 books in total or an average of about one book per undergraduate student in the Cornell population (compared to approximately 8 books, on average, for graduate students and about 13 per faculty member).”

They continued: “The library in the research university has traditionally aspired to build a collection that would satisfy any potential research need; that some portion of the collection would remain indefinitely latent has generally been accepted as the condition for meeting the needs of scholarship. What significance the Library and the University should assign to non-circulating material in today’s academic context is far from clear, however. If half of CUL’s monograph purchases of the past twenty years have circulated, is that a lot or a little? Precious resources are being spent to purchase, house, and preserve these books, but to what extent should this be regarded as misspent funds and to what extent as investment in a strategic reserve? The answer will surely vary by field and by the intended readership for particular segments of the collection. Factors such as language of publication can place distinct limits on the pool of potential users and any meaningful measure of usage must take the size of the user population into account.”

They recommend that the university monitor usage and use the resulting data to guide book purchasing decisions in the future.

Now this may make sense in theory — we can imagine a few purchasing errors, but by and large a universities libraries have to aim at some sort of comprehensiveness. Usage numbers certainly shouldn’t be ignored, but surely the librarians will be missing a vast amount of usage data if they proceed as planned. I cannot think how many more books I have consulted in the library without ever checking them out, than books I have borrowed. Unless the Cornell Library has closed shelves where in order to see a book you have to fill out a slip, as at the main New York Public Library, they will have to be be ignoring all this use. Consulting a book doesn’t have to mean reading it cover to cover. They recognize this problem and even have a name for such in-library usage: “historical browses”, a somewhat trivializing term, data for which they admit they don’t have and have thus decided to ignore. “We recognize that circulation is an imperfect surrogate for use of items in the collection.”

Frankly I think that the main advantage e-books have over print books may amount merely to the fact that their usage is easier to track. I bet many more print books are “used” in libraries than usage can track. Researchers tend to find it easier to collect references via digital files, but following up those references and reading still seems to be preferred in the print product. Not every follow up has to involve checking the book out. Maybe librarians need to invent some “historical browse” detector. One might imagine a chip in the spine which detects and records when the book is opened and when it is put back on the shelf. Or a finger-print detecting cover. Constant video-ing of the shelves might creepily collect behavior better left unseen. If librarians are going to use usage data to inform their purchasing decisions shouldn’t the onus fall on them to ensure that their data are in fact complete? It could be that some of the 55% of books “never circulated” have in fact been consulted many times more than some that were checked out. I can imagine myself heaving a sigh and saying “Oh well, this one’s so badly organized that I can’t find what I’m looking for; I’ll have to take it home and go through it tomorrow”. This would turn up as usage, whereas the better organized books I had consulted might appear never to have been looked at. In so far as these books might fall into any particular class, the purchasing policy informed by such faulty data is going to result in faulty buying decisions.

We should resist the temptation to make our libraries more efficient. They are not involved in a manufacturing process where greater efficiency is an unambiguous good. Libraries contain books and allow people to discover information. This is just inherently messy.