Thinking more about ebooks and libraries and what big publishers should do
The reluctance of most big publishers to make ebooks available through library lending is a topic of widespread attention and concern. The AAP turned a chunk of its annual meeting over to the topic and Dr. Anthony Marx, the President and CEO of the New York Public Library, used his time to volunteer his institution for experimentation to find a model for ebook lending that would work for publishers.
I had occasion to talk to a number of Big Six publishers in the middle of last year about their position on library sales. When they registered their concerns with me, some of what they had to say made a great deal of sense.
What really rang true was the fear that the consumers in an emerging ebook ecosystem would “learn” that getting “free” ebooks from libraries was just as easy as getting ebooks from retailers and paying for them. Given that all this requires is pointing your web browser in a different direction, it looked to many of the publishers like a really poor bet to enable ebook lending by libraries. Sales of ebooks to libraries isn’t a huge market so the upside is limited. And with many ebook retailers struggling to gain traction in an Amazon-dominated marketplace, the consequences of even a small loss in sales could knock players out of the game.
Withholding or limiting sales of ebooks to libraries shares an important characteristic with agency pricing. In both cases, publishers are implementing policies that they know will result in their revenue being reduced immediately in order to develop what they believe will be a stronger and more diversified distribution network for ebooks in the long run.
It’s worth making a point here. It is reasonable to argue that publishers are wrong on both counts. These policies are about influencing the future development of the channel and forecasting the impact of various policies over time, so there is not yet any data to prove they’re right or wrong. All we have are guesses based on limited and unprojectable knowledge.
But the publishers are most emphatically not engaging in “short term thinking”. They are sacrificing immediate revenue for what they believe will be a long-term gain. In that way they are like Amazon, which famously will deep-discount or loss-lead today’s sale to build a long-term customer relationship. But Amazon gets credit for long-term thinking; publishers usually don’t.
As most people know who are following the tribulations of libraries trying to stock ebooks, four of the Big Six publishers are not making any ebooks available to libraries at all (except titles already sold in the past.) Random House continues to supply all their titles to libraries as ebooks with only the “one loan at a time per copy purchased” limitation, but they have just raised the prices of those books to libraries substantially. HarperCollins was widely villified a year ago when they introduced a limitation of 26 loans per copy purchased, but this is apparently now more widely being seen as an acceptable limitation. (Random House pricing and the others’ total withdrawal from the market are making 26 loans look good!)
I accept the major premise. If it were just as easy to get ebooks from libraries as it is from retailers, over time more and more customers would migrate to the libraries. But, the more I think about it, the less I accept the notion that total withdrawal from the library market is necessary to create a clear advantage for the retailer as a destination for ebook readers. In fact, it is possible that putting ebooks into libraries, in the right ways, could increase sales at retail. And the only way for publishers to find that out is to do some controlled experimentation in that marketplace. To my knowledge, that’s not taking place.
I have two anecdotes that I think shed some light.
At a conference I organized a few years ago, we heard about a giveaway that publishers Spiegel & Grau did of a Suze Orman book through Oprah. They gave away 1.1 million unprotected PDF downloads in 33 hours on Oprah. The book sales popped immediately on Amazon, of course. But more impressive, and more important, is that the book acheived a return run of several more months on the NY Times Bestseller list. It had fallen off in October. In other words, the substantial number of giveaway copies sparked a rebirth of interest in what had already been a very successful book that appeared to have concluded its run some months before. (Of course, being on Oprah has an effect of its own whether or not there is a giveaway. I think the 1.1 million giveaways helped spur the sales, but, at the very least, they didn’t prevent the Oprah effect from taking place in a big way. The publishers had not expected anything like the commercial result they got.)
And here’s another. About two or three years ago, I was looking around for data that would tell us whether ebook sales were cannibalizing print or adding incremental units. Our friend Rick Joyce at Perseus said “they add incremental units” and he could prove it. I was skeptical, but he convinced me.
What Rick said was that Perseus had converted a number of backlist books that had an established print sales pattern into ebooks. Then they looked at what happened to sales of the overall “basket of titles”. Their print sales went up even though the ebook was selling too. Perseus attributed this to the increased awareness of the books generated by their presence in the ebook marketplace.
But perhaps of even greater significance, the beneficial effect on sales of the ebook publication increased as they went down the long tail. The deeper they went into the backlist, the greater was the lift created by the ebook publication.
It is a stretch to analogize the effect of an ebook being available in a library today with what making an ebook for sale did to print sales two years ago, but it is not a ridiculous hypothesis to test. What if a similar impact resulted from library availability? Might there be some titles you want to window, but others where it would even make sense to sell them to libraries cheaper because of a marketing effect? Could there even be some books that it would be worth giving to some libraries for free?
Both of these examples, being as they are more than two years old (and, in one case, about PDFs, which are a pretty limiting way to get an ebook file), can’t be simply taken as prima facie evidence of what would happen today. People who downloaded the Orman file and found they liked it and wanted to read the whole book would indeed have wanted to purchase a more readable copy. In 2008, that would have meant “print” for the vast majority of people. Perseus did their backlist analysis at a time when they had lots of good backlist not yet available as an ebook and not many consumers read that way. Neither of these things are true anymore. The migration of sales from print to digital is too rapid now for very many titles to maintain their recent print sales rate, let alone increase it on the back of some increased discoverability. (Overall sales might increase, but not print as a stand-alone.)
Nonetheless, these two pieces of anecdata together suggest at least the possibility that the sales the big publishers are losing by withholding from the libraries is a larger number than just the ebooks they’re not selling for loan. They may also be losing other sales that come from discoverability and library-reader-generated word-of-mouth.
It now seems to me that making ebooks available through libraries when the titles are on a downward sales trajectory at retail could generate new life for them, as the Oprah giveaway did for Suze Orman. Yes, this is windowing. If publishers did it on their big ebook titles, they’d be doing exactly what Hollywood is doing with DVDs of major movies, which are also withheld from library distribution until the theatrical and early DVD revenues have been harvested.
The big publishers I have spoken to seem most focused on keeping “friction” in the library ebook experience to approximate the inconvenience of print book borrowing, where you have to go to the library to pick up the book and then to bring it back. In fact, the imperfect interface from OverDrive already provides a good deal of that (except for Kindle loans, which bump over to Amazon and work seamlessly). The absence of any new titles from four of the Big Six may not provide “friction”, but it certainly would drive many readers to a retail channel. (In fact, without some huge change in the way publisher-library relationships operate, there will always be a much larger number of titles available from retailers than from any library.)
What publishers are correct to be worried about is the way the market could change over time. With most book readers still reading print, we don’t really know yet what the marketplace will look like in five or ten years when I’d expect the vast majority will probably read books on screens. That was part of what was behind Harper’s 26 loan limitation. It is also a principle behind Bloomsbury’s “Public Library Online” program (working in the UK for a while now and just being introduced here), by which a “shelf” of books is licensed to libraries one year at a time. (Public Library Online enables multiple users through any flash-enabled browser, but does not support even as basic a user tool as “place-holding”; each time you return to a book you would have to remember where you were.)
In general, time-limiting seems a much better strategy to me than loan-limiting, although it might not produce the same level of additional sales on bestsellers as loan-limiting.
If any big publisher asked me for an opinion about a library policy (and none has), this is what I’d say today.
1. Start immediately experimenting with “baskets” of titles, because the data on sales trends for a group of titles will be far more reliable than on any single title. If titles are put into groupings of cohorts (fiction in a genre, topical non-fiction, big author brands), you increase your chances of getting data that lends itself to interpretation that enable useful adjustments in tactics.
2. One set of experiments that should be productive would be on titles that have already had their high-volume run. Put 10 or 20 of those titles into library distribution and look at their print and ebook sales results week-by-week for the period before and then after the library release. (And promote the library release to maximize the potential impact.)
3. Look at the “make” books on an upcoming list: those that aren’t by big name authors that are already guaranteed to sell well. Split them in half. Put one half into libraries and withhold the other half. See if you can detect a library effect, positive or negative.
4. License titles for two or three years rather than limiting the number of loans. This will enable the publisher to withdraw them from library circulation in the future if the market shifts. This is a separate question from whether you allow multiple simultaneous loans. That limitation probably needs to remain (although with a loan limit like HarperCollins has applied, I don’t see why it is necessary.)
5. Explore ways for libraries to sell ebooks to patrons who discover titles through them but, for whatever reason, want to purchase them. Referral to existing retailers, with libraries getting the referral revenue, would seem like the cleanest and best way for this to happen. Libraries could sell the ebooks directly, but that approach could exacerbate the concern that library patrons would be “lost” to the retail network.
Publishers’ concerns about the impact of library lending are reasonable. But responding to that concern by simply “freezing” is not helpful to anybody and it may actually be damaging the sales of the books the publishers are trying to protect. I don’t know and the librarians don’t know what the marketplace impact will be of branded ebooks being made available through libraries, but the publishers don’t know either. It is time for all of us to start finding out.
I think big publishers have widely accepted a similarly flawed perspective about the impact of piracy. It is likely that piracy, like library loans, will have a net cannibalizing effect on some titles and a net promotional effect on others. If you accept the truth of that statement, then you should see that telling which titles are which is the most important starting point to creating policy in either case. How many houses today are consulting their marketing departments when they make their piracy policies? It would seem like a good idea.