This seems to me to suggest that free downloads don’t do much to
cannibalize actual [book] sales.
I have more (or at least longer-term) experience with this than
anyone else. Back in 1991, The New Hacker’s Dictionary
was the very first real book (like, with an ISBN) to be released
simultaneously in print and available for free download on-line. Both
of the books I’ve done since, The Cathedral and the
Bazaar and The Art of Unix Programming, have also
been released for free download at the same time they were in print.
You can easily find all three on my
Of all my books, only the very first (Portable C and Unix
Systems Programming, 1987) didn’t get webbed. It was a decent
seller, but the least successful of my books. It’s now out of print, made
technically obsolete by things that happened in the early 1990s. All
three of my other books, the ones that got webbed, have remained
continuously in print.
My four books do not a controlled experiment make, but the
thirteen years of experience with simultaneous print and Web
publication that I’ve had suggests that Web availability has boosted
the sales of the print versions tremendously. And my publishers
agree. Even in 1991 I didn’t get resistance from MIT press, and
Addison-Wesley was positively supportive of putting my most most
recent one on the Web.
I’m one of a handful of technical-book writers who publishers treat
like rock stars, because I have a large fan base and my name on a
cover will sell a book in volumes that are exceptional for its
category (for comparison my editor at AW mentions Bruce Eckel as
another). I’m not certain my experience generalizes to authors who
aren’t rock stars. On the other hand, it’s more than
possible that I’m a rock star largely because I have been
throwing my stuff on the Web since 1991. It’s even likely —
after all, I was next to an unknown when I edited The New Hacker’s
So I don’t find the InstaWife’s experience very surprising.
Webbing one’s books seems to be really effective way to build a fan
base. My impression is that people start by browsing the the on-line
versions of my books, then buy the paper copy partly for convenience
and partly as what marketers call an identity good.
An identity good is something people buy to express their tie to a
group or category they belong to or would like to belong to. People
buy The New Hacker’s Dictionary because they are, or want
to be, the kind of person they think should own a copy of it.
Here’s the causal connection: A Web version can’t be an identity
good, because it doesn’t sit on your bookshelf or your coffee table
telling everybody (and reminding you!) who you are. But Web exposure
can, I think, help turn a book with the right kind of potential into
an identity good. I suspect there is now a population of psychologists
and social workers who perceive the InstaWife’s book as an identity
good, and that (as with my stuff) that perception was either created or
strongly reinforced by web exposure.
If so, this would explain why webbing her book made the auction
price for the out-of-print paper version go up. The price of the
paper version reflects buyers’ desires to be identifiable as members
of the community of readers of the book. By making softcopy available
for download, the InstaWife enhanced the power of the paper version as
an identity token, by making it easy for a larger population to learn
the meaning of the token.
I would go so far as to predict that any book (or movie, or CD)
that functions as an identity good will tend to sell more rather than
less after Web exposure. All three of my in-print books happen to be
identity goods rather strongly, for slightly different but overlapping
populations. I suspect the InstaWife’s book has this quality too. About those
things which aren’t identity goods, I can’t say. Not enough experience.