Book Reviews: Why They No Longer Help in Book Discovery and How to Improve the Search for Good Reads


Book reviews are the key to book discovery. That's conventional wisdom. And it's rooted in what the publishing industry has been doing since Dicken's time to leverage the effectiveness of book reviews:  major literary critics and authors are asked to write book reviews that are then published in major papers with a wide circulation or literary journals with a targeted audience.

With the digital revolution and the rise of the e-book, all that is changing. Traditional publishers are losing ground to what is happening on line where book reviews are "diluted" across a myriad of e-platforms, from the biggies like Amazon's Kindle Store where customers are encouraged to write reviews to thousands of small book review blogs. New websites are springing up every day, from Read Wave to Wattpad. The latest is BookLikes that claims to be like a "new Tumblr" for authors. 

Now that wouldn't matter much if the market share of e-books was small or declining - but it is large and rising, and therefore how book reviews are handled on the Net is becoming increasingly important. Consider the size of the problem: by 2012, according to a new Bowker research, e-books had clearly "come of age" with a market share of 44% of book purchases by volume in the US and 38% in the UK.

Add to this the fact that the digital revolution has brought on a tsunami of new titles. For example, the Kindle Store has some two million titles with over 1,000 added each day.

How can a poor reader find his/her way in this tsunami of e-books and online reviews?

The short answer: he/she cannot. Traditional reference systems (like the New York Times or Granta) are over-run, indeed Granta is in crisis. Publishing Trends reported that waves of spammers were hitting e-book stores (see here), among them a clever guy going by the invented name of "Manuel Ortiz Braschi" who published over 3,000 books on the Kindle Store, including public-domain titles like Alice in Wonderland.  


Mike Essex, a Search Specialist at UK digital marketing agency Koozai tested the system by publishing a  song lyric and repeating it over and over throughout an e-book without Amazon detecting it. As he wrote in his blog: "It’s maddening. The logic of ‘the market will decide’ is flawed. How many customers have to be ripped off by shoddy content that adds no value before someone leaves a bad review? There’s no option to report a book as spam, and people can get away with rubbish content which dilutes the offering for good authors." 

That was two years ago, back in March 2011.

Has anything changed? Yes, I double-checked. Manuel Ortiz Braschi no longer exists on Amazon and where appropriate, titles he had taken over were given back to their rightful authors (like Reverend Andrew Murray's Absolute Surrender). 


Also there is a simple option to report a book spam: just contact Amazon.

So does this mean that Amazon-style online customer reviews "work" as quality control of the market? 


Not quite. There was the scandal around "sock-puppet reviews" that exploded last year when some writers were caught using family and friends and paying services for glowing reviews, not to mention the unsavory practice of writing reviews under a false name damning  rival authors books. Best selling authors have admitted to using such services, notably R.J. Ellory, the British crime writer and John Locke, the self-published author who sold one million books on Kindle in five months (now we know how he did it, it wasn't some sophisticated marketing technique like "loyalty transfer" as he claimed...).

Is the scandal over? More or less. Although Amazon went a little too far and
faced a backlash from authors when it quietly removed reviews from writers in the same genre (for examples, see here), it did move to police its review system and now it regularly catches authors even when they use different Amazon accounts.


Does this mean that all is well with the online review system? I've blogged before about how it could be improved through encouraging reviewers to link more closely the content of their reviews to the number of stars they award a book (see here). But this amounts to minor tweaking of the system. 

What would probably cause a major boost to book discovery would be a change in the search system to "kick up reviews in full view"

As of now, there are three major systems to browse for books: click "fiction" and you can get a breakdown by Best Sellers, Hot New Releases and Editors' Pick - plus a line of recommendations based on "items you own" (that works if you log yourself in). Plus all sorts of other goodies: "New and Notable", "Father's Day/Mother's Day etc Gift in Fiction", "Best Selling Fiction in Featured Categories" (that includes poetry and short stories), "Sponsored Selections", "Featured Fiction Books Rated 4 Stars and Higher" and more.

What is evident is that the bestseller notion is worked to death: you have the current top 100 in titles that are paid for and those that go free, and those that stay the longest in the "top 100" category. As I write, Dan Brown's Inferno is number one, it's been up there for 103 days.

Pretty sophisticated! Yet, there are major drawbacks to the system. 

First, "items I own" are not a good indication of what I'm looking for: I could feel I want something different (I often do and I bet you do too) and Amazon cannot know this - plus it recommends books that one already owns (the algorithm needs a little improvement there...).

Second, it's difficult to get out of the Top 100. It's not intuitive, you have to click many times to move away from bestseller lists. Then when you do, you find you can rank the titles by "new and popular" (meaning highest sales), by price (gives you no idea of the book's quality) and by publication date (who cares?) and also, oh surprise, by "average customer review". Alas, the meaning of that phrase is obscure but, upon verification of a few, I noticed that the top title, while it had fewer reviews than, say, the third-listed title, it did not have any one-star or two star-reviews like the other did - which suggests that this is a star-weighted system of ranking.

Three, Editors' Pick that includes things like the "the Best Books of the Month" and "Kindle Select 25: Exciting Books This Week", is a black box: who decides, who says what is best and exciting? Presumably the titles are picked by Amazon's own publishing imprints. 

More generally, all those so-called "featured" and "sponsored" categories do not have any clear definition and Amazon does not provide one. For example, if you dig into the list of books "Rated 4 Stars and Higher", you'll find to your surprise (I was!) that there are only 6 titles in there and all of them except one not yet published. They are in fact reviewed in advance, presumably through Amazon's Vine Program.  

In short, there are two major navigating systems available to readers looking for new reads in the Kindle Store (or any other e-book platform). One is sales rankings, the famous "top 100", and it remains the major one and most talked about: that's the one authors compete to enter in, that's the one reported in the press. The problem? Sales give you no intimation of how good these books are, they don't help in book discovery.  If anything, since success breeds success, there is a tendency to always find the same authors in the top 100.

Is the other major search system proposed by Amazon, the one based on "average customer review" of help in book discovery? One problem, as mentioned above is that the criteria used to define the category is not clear. The second and bigger problem is that the top 100 slots are solidly occupied by books with reviews in the thousands - clearly titles that have been "popular" for quite a while and are selling accordingly. In a sense, while the books are not ranked in the same order as the top 100 in "paid in Kindle Store", the list is a duplicate and doesn't help in book discovery. It's not cool.

Plus this list has some serious anomalies: for example, Jodi Picoult's The Storyteller has over 1,700 reviews, yet it ranked #200 on the "average customer reviews" list but, surprise, it stood at #53 in Paid in Kindle Store. Why is it so far down the "reviews" list? Maybe because it has a very high number of bad reviews (31 one-star and 56 two-stars) or maybe not.

In short, this approach to book discovery is a dead-end. 

What is needed is a search system that would identify high quality books in a "cool" top 100 list.

How to get there given the limitations of the present system?  You could imagine a cut off point and only list books that made the cut. And of course rank them. But how would you define that cut-off point?

Could outside forces - say the members of Goodreads or other book reading clubs - be somehow drawn in to help set up this list? 

Could Amazon use other parameters than book reviews?

Any ideas?

My sense is that Amazon's book reviews system doesn't catch new "good reads" as they are published. The Vine Program as I blogged before (see here) is an essential instrument but it could probably be upgraded and the best reviewers turned into some sort of vetting committee. I know self-publishers hate gatekeepers and that is what traditional publishers are most criticized for. 

Yet something needs to be done to ensure book discovery in the digital age: by itself, contrary to what is commonly believed, the cream will not rise to the top. 

My question to you: If reviews are no longer any guide to quality reads, are literary prizes - the Man Booker, the Pulitzer etc the answer?


Enhanced by Zemanta
2 comments

Popular posts from this blog

How One of the Internet's Founders Sees the Future

Rome has Become a Mess!

AUTHOREA: A STARTUP FOR SCIENTISTS TO SHARE AND ADVANCE RESEARCH