Kindle Review – Kindle Fire Review, Kindle 4 Review

Andrea Belz pens a very optimistic article about Publishers’ prospects in the brave new world of Publishing. She feels that Publishers are well placed because they have branding and trust and will function as ‘Filters’ that tell readers what books are worth reading.

Bringing up the need for Filtering is a very smart move and since Filtering is so important it’s what this post will dissect.

Do we need Filtering in an era of eBooks and eReaders?

That’s almost a rhetorical question. We need filtering just to decide which free book offer to sample first. For eBooks in general we desperately need good filtering.

Who’s going to do the Filtering? Who could?

Publishers think they are sitting pretty because they have brand recognition and finances and are used to their role as gatekeepers. However, there are actually multiple contenders vying to take over the task of filtering books for readers -

1. Algorithms – Recommendation Engines from Amazon and Google and Apple’s Genius Recommendation Engine.  
2. Really Intricate Algorithms – Perhaps the equivalent of a Pandora like system that dissects each book by genre, tone, language, emotion, and feel.  
3. Users via Reviews – This is already being done at Amazon. Reviews are usually the best data point to decide what free book to go after first. 
4. Users via Purchases – The bestseller lists on Amazon are hugely important and bestseller lists will continue to be extremely important.
5. Publishers – They don’t realize that they are arguably the most expensive Filters that could be set-up. How can you compare free user reviews with Publishers that want a 75% cut?
6. Book Reviewers – People go to reviewers before they go to Publishers.
7. Authors – The real branding is with authors.
8. Book Blogs – Sites that cater to book lovers might be able to take over filtering.

You have a lot of different aspects that come into play – the premium the ‘filter’ entity charges, trust, branding, ability to scale, the human touch, and lots of others variables.

There are combinations of these variables and situations that favor Publishers. However, Publishers have very high costs. A book review site that wants a 5% cut of the book’s price easily beats a Publisher who wants 45%.

Could Crowdsourcing Work? Is the intelligence of the crowds harvestable?

The most elegant solution would be to remove all the middlemen and have readers help each other.

We already see this at Amazon where user reviews are an extremely important value-add feature. They’re all reviews shared by users – for free. It’s hard to get a more genuine perspective on what a book is like and it’s hard to get as unbiased an opinion as that of a fellow reader.

There are lots of ways to harvest the wisdom of the crowds - reviews, ratings, personal recommendations, status updates from users, purchase history, and more. With eBooks we not only have what users say we also have information on what they actually bought, what path they took to the purchase, what books they chose between, and what other books they bought in the past.

Publishers should keep in mind that readers are happy to do lots of filtering for free and might even be able to do a better job than Publishers.

Would Algorithms go well with Crowd Intelligence?

 Actually, they’d go very well.

Readers and the Crowd would provide a lot of reviews and recommendations and valuable data on user behavior. Then algorithms would step in and use all the data and reviews to extrapolate and build up a whole model – what readers like you liked, what the structure of books you like is, which new books match your tastes, and a whole lot more.

Perhaps the most fascinating possibility is a Pandora for Books. Consider this snippet from the linked article -

So why isn’t there a service like this for books, something that looks deeper than customer buying habits at Amazon or broad genre categories?  Well, there is!  It’s called Book Lamp and it tries to match readers who like a given book with other similar books based upon not only genre but also things like tone, tense, perspective, action, description, and dialog.

Problem is, Book Lamp now only has a few hundred titles in its database …

Once we have an algorithm that analyzes tone, dialog, language, structure, and grammar (in addition to the usual suspects like genre and past purchases) we could build up a recommendation system that would not only suggest a book for you it would tell you how much you would like it, the reasons why you’d like it, and list the 10 people in the world with an identical taste to yours and what they thought of it.

User behavior and users with shared interests are all that is needed – With enough data we could predict the entire reading history of a reader and also most of her/his reading future. Amazon, Apple and Sony are well placed to do this since they have all this user data and user history. So are Google.

Filtering is simply Search for Books

It’s amusing to think there was a time when people decided where to go on the Internet based on human edited directories. That’s exactly what used to happen though – until algorithmic search and search engines came along.

We’re in a very similar situation with Books – We wait for people to tell us what to read.

It’s gradually beginning to shift though. We might not be that far from a time when we just type in ‘a book that chills me to the bone’ and are provided with half a dozen recommendations. It’d be a system that would be even more advanced than search because it would know what we read, what we bought, what we considered and didn’t buy, what we rated books, and which books we recommended to friends.

If a Pandora for Books were put together we would take all the user behavior and shared interests a step further – we’d be able to tell what were the key characteristics of a book that appealed to each reader. It’d be like having a recipe for what most appeals to every reader’s palate. We could actually have authors writing books based on the qualities their audience (or prospective audience) most treasures. We don’t just mean genre – it’d be dozens of things like amount of character development, pace of the book, amount of twist in the plot, the type of ending, and other important and intricate details.

Publishers are mistaken if they think the Search for Books is going to be human-powered

Unless you are leveraging the free contributions of readers there is no way to match an algorithm on efficiency and cost effectiveness.

Publishers aren’t being able to survive on $10 eBooks – Are they really going to be able to compete with algorithms that cost a few cents per user book recommendation?

The only question left with filtering is whether it’s going to be algorithms combined with crowdsourcing or just algorithms. The platforms i.e. Apple and Amazon are very well placed to take over filtering and all of Publishing. You have reviews and ratings and bestseller lists and other readers and a captive audience – How is anyone supposed to compete with all of that?

Filed under: publishing Tagged: | filter for books, future of publishing