Company Blog
Another Open Letter to Amazon
by Jason on 05/29/2010
Today I came across an Open Letter to Amazon in which author John Betcher proposes how Amazon could create an "indie book index" to help buyers know what's good and what isn't among independently published and self-published books.
The problem Mr. Betcher addresses is a serious one. It's not easy for independent publishers to get trusted, well-known book reviewers to review their publications. It's even harder for self-published authors. Yet, when shopping online, reviews are one of the chief sources of quality assurance customers look for.
I encourage you to read the full text of Betcher's open letter, but in short his proposal amounts to Amazon a) allowing customer reviews on indie and self-published books (note: they already do allow customer reviews on independent publishers' books; I can't personally speak to how they treat self-published books), and b) developing some kind of algorithms it can use to filter out spammy ratings for times when a self-published author rallies fifty of his closest friends and family to rate up his book.
It's a good idea, and would certainly help level the playing field for indie and self-published books versus books from the big-name publishers.
However, based on my experience running this website, I see a different solution.
I don't think Amazon should necessarily be in the business of trying to automatically filter out spam from customer reviews. People understand that those represent a certain sort of "Wild West" of information, and know to take it with a big grain of salt.
The real problem Betcher is encouraging Amazon to solve is how to give customers confidence that a book doesn't suck.
For that, a stamp of approval from a reviewer who has experience, and thus credibility, is key. So my suggestion would be that Amazon create a system whereby indie and self-published authors could earn those, rather than trying to police the Wild West of customer reviews. I mean, sure I'd love it if I could just pick up the phone and get Kirkus to review everything in Long Tale Press's catalog. That would be awesome. But it's not the only way to provide that stamp of approval.
If I were Jeff Bezos, here's how I would do it:
Data-mine the existing set of customer reviews to find customers who are already in the habit of reviewing a lot of books, and doing so in a substantive and unbiased way. To some extent Amazon already does this (and who knows what they do that we don't yet know about!)
Invite those people to join a pool of "Amazon Approved" reviewers. These people could submit reviews that would appear above the fold on Amazon's book pages, in the space where you typically see reviews from Publishers Weekly and other well-known reviewers.
Data mine the tastes and preferences of those special, blessed reviewers, so as to send books to the reviewers who are best able to judge that book's merits (e.g. don't send a Sci-fi book to someone who only reads Harlequin romances).
When indie and self-published books reach a certain threshhold of sales, send a free copy of the book to an appropriate person from the review pool. Done.
The idea here is that Amazon could use the data it already has to kick-start a community of trusted reviewers large enough to supply the reviewing needs of books that in fact deserve a review. Because not all of them will.
Amazon is still going to have to balance the workload: there will always be more self-pub and indie books out there than any reasonable pool of reviewers can give proper attention to. But, if you know the slushpile like I do (and since we put our slushpile up for all to see, you do), you'll know just as well as I do that most of those books really don't deserve any particular attention.
I think this model would work better than Betcher's proposed system for a few reasons. First, it would be a lot easier to implement. Trust me, I was a programmer in a past life, so I have a good sense for evaluating the technical feasibility of ideas like these. Betcher's idea pre-supposes that we can accurately teach computers to infer the intentions of people submitting customer reviews. That's a hard, hard problem for a computer to solve. My proposal doesn't do that, but relies on the ability to merely detect people who are good at doing reviews. These people can be vetted by hand, if necessary, by spot-checking their reviews.
Second, my proposal is an extension of an existing model that already works. The whole reason why people trust reviews from Kirkus or Publishers Weekly or whomever is precisely because not just anybody can join. It is an intentionally exclusive club. Customer reviews have their place (and I often make my final buy/no-buy decision based on them), but I don't think they're the solution to the problem Betcher is pointing out. Rather, let's have Amazon tap its vast, existing community of book lovers to create its own intentionally exclusive club, and turn them to the task of evaluating the tidal wave of indie and self-published books.
Company Blog