Bad Math Among eBook Enthusiasts

The Amazon Kindle has excited a lot of comment from people who have long wanted portable electronic access to books. Amazon has put together a lot of the pieces that makes this holy grail seem reachable, even if not yet truly achieved. But in reading the commentary of some of the enthusiastic boosters of eBooks, I’m struck by just how much wishful thinking they display. For example, in an interview on Teleread with Marie Campbell of MarketIntellNow, Marie reports on a poll of 5000 “random” (actually self-selected) web surfers, and concludes: “Lower e-book prices, not gizmos like the Amazon Kindle, will be the big spur for book sales.”

I agree with Marie’s fundamental assertion that handheld devices are nice to have, not need to have — the real breakthrough is just making electronic copies available at a fair price. Eventually, those electronic copies will find their way to whatever devices we are currently using, whether it’s a PC or a handheld or some other future device.
Our Safari Books Online joint venture with Pearson, which delivers subscription access to thousands of computer and business books (admittedly an early adopter audience with a need for professional content), generates more revenue than is normally reported for the entire downloadable ebook business. For my company, O’Reilly Media, Safari is now our third largest reseller, behind only Amazon and Barnes & Noble. This didn’t take a handheld device (though handheld devices will certainly enlarge the market.) It just required making the content available with a business model and price that people were willing to pay.

But Marie’s poll goes on to conclude that “over half of people surveyed expected e-book prices to be $5 or less and 1 out of every 5 expected the price to be $2.50 or less.” She says:

We believe that the Publishing Industry will very quickly discover that they’re blessed with ELASTICITY. That is, the lower the prices of e-books up to a point, the more net revenue they drive (thus the cannibalization effect on traditional book sales will be overcome). E-books may start around $10.00 each, but come down in the 2008-09 timeframe and approach $5.00.

It’s true that new price points can sometimes attract new readers — that’s what happened with the rise of the paperback. But note that as paperbacks became the dominant format, outselling hardbacks, prices rose substantially for both paperbacks AND hardbacks. They didn’t keep falling. So if prices fall to $5 or less, as predicted, you can equally bet that they will rise if and when the electronic format becomes dominant.

What’s more, I think that the idea that there’s sufficient unmet demand to justify radical price cuts is totally wrongheaded. Unlike music, which is quickly consumed (a song takes 3 to 4 minutes to listen to, and price elasticity does have an impact on whether you try a new song or listen to an old one again), many types of books require a substantial time commitment, and having more books available more cheaply doesn’t mean any more books read. Regular readers already often have huge piles of unread books, as we end up buying more than we have time for. Time, not price, is the limiting factor.

And as for the kind of books that you don’t read from beginning to end, but just use to do a job like looking up information, or learning something new, the “all you can eat” subscription model may be more appropriate. With Safari, we’ve increasingly moved from a “bookshelf” model (in which you put books on a bookshelf and can only swap at month end) to an all you can eat model, because we’ve discovered that people consume about the same amount of content regardless of how much you make available. All you can eat pricing lets people take what they need from more books, but it doesn’t increase the total amount of content they consume. It merely changes the distribution, and in particular, favors the long tail over the head. (See my post from last year, Long Tail Evidence from Safari and Google Book Search, for details on how a searchable database changes the distribution of content that is viewed.)

Cable TV provides another argument for the subscription model for content types that, unlike music, require a time commitment once a choice has been made. “Channels” and subscription packages generate far more revenue than pay-per-view, which is the equivalent to the downloadable ebook.

Ultimately, I believe that we’ll see a variety of models for electronic content — just as we do in print: single copy sales, subscription, advertising, and even wacky hybrid models. I wrote about this back in 1995, in an article called Publishing Models for Internet Commerce.

But in understanding which model to apply, you have to do the math!

Marie goes on to say:

E-books will start sans ads, at prices of about $10/book. But as the $5.00 point-of-pain for publishers is reached, hybrid models will begin to flourish. And some outriders- likely VC-backed- will stake out the “free books supported by advertising” segment and drive earnings for all to see.

Marie obviously hasn’t read Jeremy Liew’s eye-opening post Three Ways to Build an Online Media Business to $50 Million in Revenue, which dissects just how many page views you need to drive to get to only $50 million in revenue — the size of a mid-sized publisher. Short answer: way more than most people ever imagine.

Obviously, the advertising model works famously for some kinds of content. But publishers are kidding themselves if they think that advertising will replace the revenue generated by current book sales. Current CPM (cost per thousand) rates for advertising range from $1 (the vast majority) to $20 at the high end, very targeted, high-value audiences. So if you have a 200 page book that sells 20,000 copies, generating 400,000 page views (assuming all 20,000 people read the entire book), you might generate a few hundred dollars from advertising even with an ad on every page. [Update: a number of people have pointed out my own bad math: that’s 4 million page views, resulting in a few thousand dollars of ad revenue. But that doesn’t change the fundamental point.] And that’s before you develop your ad-serving technology or pay someone else a slice for handling the problem. Publishers and authors have to get a LOT more readers to bring you up to the level of revenue you get today from a printed book.

Advertising works for aggregators like portals and search engines because they can amass billions of page views, often on user-generated (or computer-generated) content for which they pay nothing.

I’m sure there will be ad supported books, and ads as a supplement to other revenue streams, but it isn’t going to be enough to support publishing as we know it.

I see wishful thinking in other projections as well.
over In Business Week, David Kiley writes:

If I sell my Kindle book to a reader for $9.99, he has saved perhaps $20 on the price of a hardcover book. Let’s say I sell 40,000 copies of the book, and further assume that I can get at least 10,000 of the buyers to subscribe to periodically updated chapters and podcasts, or perhaps a blog, for an additional $9.99. As an author, I would like to control that end of the revenue stream, which I don’t need a big publisher for anyway. That’s another $100,000 in revenue to me, minus my costs and Amazon’s cut.

Leaving out the part that selling 40,000 copies of a book is very good to begin with, and that “Amazon’s cut” for author-published content is 65%, reducing that imagined $100,000 to $35,000, what basis does David have for assuming that 25% of his buyers will pay to subscribe to updates? In the late 90’s, we offered a product line called the CD bookshelf, collections of related books on CD that were sold through bookstores. We offered the opportunity for people to register their purchase so they could get updates to the books, and were shocked to find how few people took us up on the offer. We sold tens of thousands of CD bookshelves, and ended up with hundreds of people registering for updates. Anyone who’s ever built a subscription business knows that it takes a large investment to build up a subscriber base that produces a significant revenue stream.

Evidence from related fields like free software downloads similarly suggests that you’re lucky if a fraction of 1% converts to a paying customer. Or take Wikipedia. ClickZ claimed they had 189 million unique visitors in January 2007, and I’ve seen estimates that they get 7 Billion page views a month, yet their recent request for donations, which appears as an ad at the top of every article, has so far resulted in little more than 33,000 donations. (Just check the header of any Wikipedia article for the current tally.)

There are a lot of good ways to create user generated content that can produce significant advertising revenue streams. And there are lots of good ways to build a subscription business, just as there are lots of good ways to make a success out of selling information goods by the piece. But none of it is easy, and none of it happens automatically. Finding the right price to expand the market is also hard, and there are clear limits to the kind of growth you can get with radically lower prices.

My advice to publishers and authors is this: figure out what it costs to produce what you sell, estimate what kind of volume you’ll be able to achieve using the best available data, and then set your prices at a level that will deliver a reasonable profit from your efforts. Sound familiar? That’s what you do in business today. Don’t expect any suspension of the law of gravity. Leave that to the subprime folks, who followed on the heels of the dotcommers in coming up with new math that ultimately didn’t make any sense.