Can a computer program predict a bestseller?

[Brian Clegg]

I've just reviewed The Bestseller Code, a book describing the way that a set of computer algorithms seems able to successfully predict the bestseller status of many of the occupants of the NY Times bestseller list. I think there are some flaws, but it's interesting for authors, and there are a few tips if you want to improve your chances of getting on that list... (my books fail most of them). See the review.

 

[Paul Whybrow]

What I find interesting about this book, is the way that it's been marketed. For a non-fiction title, it's received a massive amount of publicity pre-publication—it's not officially published until the 20th September, but articles about it have appeared in the publishing trade press and newspapers for months.

As I've commented before, there are people making more money from support services for wannabee writers, than they ever would from writing their own novels. These can be editing, critique and writing services, or money-spinning competitions that add to the wealth of the organiser, but do little for the fame of the winner. There are some bizarre leeches out there: the 'five-times-great niece of Jane Austen' is currently offering a Jane Austen Writing Masterclass—only £35 for two hours of complete tosh!

Just as store owners in gold rush times, made more money selling equipment and clothing to miners, so there are lots of clever folk around flogging universal panaceas to desperate writers, who hope to learn the secrets of success.

When we think of best-selling books, it's novels that come to mind. But, there are nonfiction titles which have sold millions. Think and Grow Rich, by Napoleon Hill has sold 100,000,000 copies since it was published in 1937.

I'm sure that The Bestseller Code will become a bestseller itself—people want instant answers.

Incidentally, Brian, you rather beg the question by saying, in your review, 'Taking a quick skim through the top 100 books selected by the analysis, there are perhaps three I would consider reading.'

Which three?

 

[Emurelda]

I don't remember where I read the following; but there's an incident when someone submitted their 'formulated best seller' based on their own analysis of various parameters. He basically averaged it up and applied it to his own fictional concoction. Was rejected. He responded insisting that it's a best selling formula and fool-proof. But the agent/publisher didn't connect with it and was astounded that anyone would think that there's a formula for this.

People would like quick answers and easy solutions but life doesn't work that way.

I think there are more factors/parameters than merely the content of the book itself. The marketing push made for it to be spread has a huge influence. Organic vs artificial marketing may also be an affect. The number of 'connectors' that share it ad convince others to buy/read it (connectors are those people who will shout about all things - check how many connectors you have in your network). For info on connector read Tipping Point. It could be that one small thing that tips your book into the market because its timing is just right.

Many times it's luck but the harder we work at it the luckier we get. Networking with connectors and improving the writing skill and creative content must all go hand in hand to create a best seller imo.

 

[Brian Clegg]

Incidentally, Brian, you rather beg the question by saying, in your review, 'Taking a quick skim through the top 100 books selected by the analysis, there are perhaps three I would consider reading.'

Which three?

I haven't got the book any more, so I can't check the list, but there were a lot of thrillers and books primarily aimed at women. I think at least one of my three was a Stephen King.

 

[Paul Whybrow]

One influential aspect of publicising a novel is the use of 'tags'. I learned about tags when I began self-publishing my short stories, novellas and poetry on Smashwords. Mark Coker, the founder of Smashwords, offers a lot of good advice in his free ebooks on style, marketing and publishing.

Tags are one-word descriptions that apply to your novel, which readers use to find ebooks of interest to them. You may be going for a traditional publishing deal, but it's still worth thinking about how your story would be sold using tags; they are a simplified version of the blurb that appears on the back cover—keywords that stand out, grabbing the reader's attention.

Thus, my latest psychological thriller would have the tags:

Murder Mummification Stone Baby Poison Madness Promiscuity Deceit Beast of Bodmin Moor Sex Lives Urban Legend Ghosts Farming

Readers browsing for ebooks can search for stories with these themes. Tags are a useful thing to bear in mind as I write, for they may help someone find my novels, whether they are online or jostling for attention on a bookshop shelf—the dreaded process of discoverability.

Just think, what someone asks you when you tell them you've written a book: 'what's it about?'

Tags are the shorthand answer.

 

[Patricia D]

I predict The Bestseller Code will be a bestseller ;-)

 

[Patricia D]

Here's a review from The New Yorker - and IMO, getting a review in The New Yorker is a good step toward having a best seller.
“The Bestseller Code” Tells Us What We Already Know - The New Yorker

 

[James Marinero]

I think it's 20% writing ability and 80% marketing. You can have the best mousetrap etc...Here's a bit more on textual prediction: machine-learning-and-bestseller-prediction-more-than-words-can-say

 

[Richard Sutton]

I think it's 20% writing ability and 80% marketing. You can have the best mousetrap etc...Here's a bit more on textual prediction: machine-learning-and-bestseller-prediction-more-than-words-can-say

and LUCK. They haven't yet figured out how to code for that, have they?

 

[Quillwitch]

I've just reviewed The Bestseller Code, a book describing the way that a set of computer algorithms seems able to successfully predict the bestseller status of many of the occupants of the NY Times bestseller list. I think there are some flaws, but it's interesting for authors, and there are a few tips if you want to improve your chances of getting on that list... (my books fail most of them). See the review.

It looks like i´m going to have to read this one! If anything to see if I am best seller material. Not that I would change anything.

 

[Quillwitch]

One influential aspect of publicising a novel is the use of 'tags'. I learned about tags when I began self-publishing my short stories, novellas and poetry on Smashwords. Mark Coker, the founder of Smashwords, offers a lot of good advice in his free ebooks on style, marketing and publishing.

Tags are one-word descriptions that apply to your novel, which readers use to find ebooks of interest to them. You may be going for a traditional publishing deal, but it's still worth thinking about how your story would be sold using tags; they are a simplified version of the blurb that appears on the back cover—keywords that stand out, grabbing the reader's attention.

Thus, my latest psychological thriller would have the tags:

Murder Mummification Stone Baby Poison Madness Promiscuity Deceit Beast of Bodmin Moor Sex Lives Urban Legend Ghosts Farming

Readers browsing for ebooks can search for stories with these themes. Tags are a useful thing to bear in mind as I write, for they may help someone find my novels, whether they are online or jostling for attention on a bookshop shelf—the dreaded process of discoverability.

Just think, what someone asks you when you tell them you've written a book: 'what's it about?'

Tags are the shorthand answer.

Nice tags! They actually work for me. I´m interested in reading your book!

 

[James Marinero]

and LUCK. They haven't yet figured out how to code for that, have they?

I think it was Arnold Palmer who said 'It's strange, the more I practice the better I seem to get'!