Over on his blog, Neil Gaiman reports on the results of Harper Collins’ read American Gods for free online promotion:

The Browse Inside Full Access promotion of American Gods drove 85 thousand visitors to our site to view 3.8 Million pages of the book (an average of 46 pages per person). On average, visitors spent over 15 minutes reading the book.

The Indies [ie. independent booksellers — Neil] are the only sales channel where we have confidence that incremental sales were driven by this promotion. In the Bookscan data reported for Independents we see a marked increase in weekly sales across all of Neil’s books, not just American Gods during the time of the contest and promotion. Following the promotion, sales returned to pre-promotion levels.

Through an online survey, we know that 44% of fans enjoyed this browsing experience and 56% did not. Some of Neil’s fans expressed frustration with the Browse Inside tool for reading through a whole book. (This poor result is partially due to two problems which were fixed soon after the initial launch – mistaken redirect to the Flash-based reader and slow image load time)

Free sampling works…who knew?


  1. Yes, but this was done on a book that had initially come out in 2001, which means that the obvious sales to be gotten — folks buying it because they gotta have everything Gaiman, folks buying it because it just got a good review, just poked its head on the bet-seller list, because it finally came out in the cheaper paperback form they were waiting for — had already been gotten, At that point, there aren’t that many sales (relatively speaking) to lose by putting out the samples.

    Whereas if this had been put up when AG was just coming out, there might well have been many folks who would otherwise have bought the book whom the sampling would turn away. They might have read a sample online and decided that they did not like the book (which to make clear is not my comment on the book – I enjoyed it, and read the entire thing out loud to my wife. Not my favorite Gaiman, but a worthwhile read.) Or they might have read the entire thing online and decided that, while they liked it just fine, they’d now rad it and didn’t really feel the need to own a copy.

    I’m not saying posting the whole thing from day 1 wouldn’t have improved sales, merely that this test doesn’t show that. It doesn’t even show that this test increased sales on the whole – it may have just accelerated sales to people who would’ve gotten around to it eventually (although even doing that is of value to a publisher.)

  2. Read the whole post on the original link, Nat. I think the test definitely shows that it improved sales. What was interesting was that during the time of the Free American Gods sales of *all my titles* went up 40%, stayed there, and declined to normal when it was done; and there was nothing else going on to support that. (I didn’t actually post about copies of American Gods directly sold online through the buy it now link [about 1600], as I wasn’t really interested in that — what interested me was that sales of all books increased in regular, non-chain book shops.)

  3. I don’t know that Nat is trying to discount the effect it had. I believe he’s saying that it is worth pointing out that it was put up on the ‘net after the initial run of the already converted buying their copies. He’s saying that it might be best to look at the timing of the free offering instead of just looking at it as the effect of a free offering. If you change one of the variables from that successful experiment, you might not get the same results. That, from what I can tell, is the caution Nat was trying to give. I could be wrong, though.

  4. JWH: That was, indeed, my primary point. But Neil is also right in that in my final sentence, I suggested that it may not have increased sales on the whole – but to clarify, I meant “over time”, that it might have accelerated purchases that would otherwise have taken place over the next year, say, and that “back to normal” that we’re now seeing may be some mix of the positive sales effect lingering mixed with the drain coming from those accelerated sales.

    But probably not.

    Really, it was more a reflection on the problems of identifying definite effect in the very-non-laboratory conditions of the world of sales… but results like that are strong enough that it’s something that should be repeated until it no longer works.

  5. But what fascinates me — and is definitely something I should see if they put in the survey (and if they didn’t, we’ll make sure is in the next) — is where you’re getting your new readers from. Because the pool of existing readers is tiny and the pool of potential readers is vast.

    And while some of the readers of American Gods online came from my website, most were sent in from places like the New York Times and Boing Boing who wrote about the promotion.

    People don’t find new authors by buying things. Not normally (this from getting show of hands checks at conventions and signings and events). The majority of them find their favourite authors from being lent a book, finding it, someone saying “here, read this”. And then they buy in paperback. And then one day they spring for the hardback of the new book.

    I’d hazard that most of what we were getting was people who’d heard of me but not read anything by me trying out some or all of American Gods, finding they liked it, and then going out and buying books by me.