My notebook has, on its first page, a little book discovery concept that I doubt I’ll ever have the time to explore in reality. Ignoring the current publishing battles for a moment, let’s look at how someone with the wherewithall - and a suitably chunky database of titles to call on, whether that’s Amazon (the obvious) or someone else - could have another crack at book discovery. That notebook page is titled “Bookhoppr”.

How it would work:

When you finish a book and enjoy it (that part is key), you go into your Bookhoppr app and mark it as read. By scanning the barcode, or entering the ISBN or ASIN, or title/author matching, or through direct pipe from your reader app of choice (“mark as read on Bookhoppr”). If you know what you’re going to move on to, you stop there. If you’re not going to pick something else up straight away, you stop there. If you’re not sure what to read next, or it’s been a while and you’re wondering what to pick up now, you tap on the big “tell me what to read next” button.

The system looks at what you’ve just read, and what you read before that, and before that - all the hops from book to book you’ve made - and compares them with other users. It takes the closest matches - let’s say the statistically most similar 1% for a touch of variation in results - and based on what those people read next from the book you last finished, it gives you one suggestion at random that doesn’t already appear in your own hops. If you don’t own it, it’ll give you a buy/sample/see reviews link. If you don’t like that suggestion, you can tell it to try again.

That’s it. Maybe throw in an extra button: “Make me different”, that gives you a rec based on the statistically least similar 1%, or books that no one read next. Maybe too an algorithmic backup for when there are no hop-based recs. But basically very limited functionality. Tell it you’ve read something and it’ll tell you what to go for next. No reviews, no attempt at being a social network because places for those already exist. Not even a “load in my existing bookshelf”, except perhaps as a way of avoiding duplicate recs, though you could certainly have simple profile information showing what you’d read. It’d be easy enough to throw in a “Tweet this!”/”Facebook this!” option whenever you use the system for a recommendation, and an API to allow other apps to interact with your read/rec flow, but otherwise keep all of that side of things to those places that deal with it already. Stick to one thing and do it well.

What it would need:

First off, a huge goddamn database, carrying all formats and able to match each to one unified “this is all the same book” entry regardless of ISBN/ASIN. Amazon’s must be just about the best for this. I can’t remember if there’s access Amazon’s product database via some kind of API, but if not, well, this is probably something only the Big A could pull off. Even other vendors with databases of their own would struggle to include everything, particularly self-published work with only an ASIN. This is by far the biggest hurdle as without this the whole thing falls to pieces.

Second, a button for “add a book to the database” for those times when even whatever one you’re using doesn’t have what you’ve read.

Third, some people able to slowly sift those additions for duplicates and roll entries in together where needed.

Fourth, it’d have to appear as a mobile app on both iOS and Android from the very beginning. That app should ideally also have a barcode scanner (or ISBN number reader even), similar to that in the eBay app, to allow easy tagging of hardcopy books.

Fifth, and this is the second biggest hurdle, it would need a big userbase to be viable. It’s dependent on there being enough user entries in the Bookhoppr list to produce recs from a wide variety of requests. And it would have to be able to do so at launch. That means a really big alpha/beta.

Sixth, it would need to be both simple and fun to use. You could, if you were so inclined, gamify the experience slightly as with Foursquare. (“Jane Q Public has just become George R. R. Martin’s biggest reader!” etc.) If it’s not quick, simple, and fun, it’s dead. If you have to jump through hoops or fill out endless text fields whenever you’ve read a book, forget it.

Seventh, it would need to be quick and simple to maintain and manage users and the database at the back end. Having peered into the murky depths of a much smaller project back with 3NJ, I know how vital that is.

Eighth, you’d need either a vast pot of cash or some way of monetizing what would be a chunky thing to organize and maintain. Ads are annoying and an unreliable income source. Offering a hobbled free version and a full-spread paid version would be shitty to the user. Perhaps run a limit on entries/recs with a very, very low barrier to clear and the equivalent of a profile badge or a little bit of glitz to say thanks for supporting it. Every thirty books (or 30, then 50, then 100 to avoid gouging long-time users) you have to pay a buck in-app, something like that. Enough that you’d have to be a regular user (and presumably a fan of the service) to hit it, and a small enough amount not to seem grasping.

Anything else I’m missing?