I mentioned The Best Seller Code in my recent post of August 4th, where I commented on a review by Sandra Elliot for The Florida Writer.  Now, having read the book, I can give you my own reactions.

First, let me say that it is a ‘must read’ for aspiring novelists, not because it reveals all the secrets of creating a bestseller (which it doesn’t), but because it will give you insights into your own writing’s weaker points.  (Assuming that there are a few.)

One aspect of the book that I found frustrating at the outset was that there was no discussion about how the ‘almost five thousand ‘ novels which were read by computer were selected.  Five hundred to these (10%) were best sellers.  Presumably all genres were represented, but in what sort of distribution?  Equal balance of male and female writers?  How about the age and background of the authors?  (There are comments on the back grounds of best-selling authors.)  What about the authors’ nationalities?  (Although all are presumably English-speaking.)  There was no mention of the age distribution of the novels, although all of the bestsellers mentioned are recent novels.  To what extent do readers’ tastes change over time?  How about the type of publisher (traditional vs indie) and the marketing budget?

There are a number of examples of the characteristics of books which tend to make them best sellers, or not, and these, of course are helpful.  But the authors admit that their computer model is only 80% accurate in predicting whether a novel will be a bestseller.  The methodology of the authors’ research used three different mapping algorithms to compare hundreds of dimensions in ‘space’.  One dimension, for example, is the use of the word ‘very’.  It turns our that authors who use ‘very’ frequently in their text are less likely to produce bestsellers.  Particular dimensions may be quite influential in predicting bestsellers.  An example is ‘human closeness’.  The computer reads the text looking for words and arrangement of words which mean that the author is writing about human closeness.  It turns out that Fifty Shades of Grey was not a best seller because of its sexual content, but because of its human closeness.

The computer was 71% accurate in identifying the gender of the author.  Three genres that have difficulty achieving bestseller status are romance, science fiction and fantasy.

Some of the dimensions which contribute to good public acceptance include: emotional cycles; active, rather than passive characters; characters who need rather than wish for; author’s distinctive style (J K Rowling’s first incognito novel was recognised not by its subject but by her style).

Topics that readers like include: marriage, death, taxes (really), modern technology, funerals, guns, school, work, doctors, presidents, kids, moms, and the media.  Less popular subjects are: sex (except in a small erotic genre), big emotions, wheeling and dealing, existential or philosophical sojourns, dinner parties.

For me, the chapter on style was particularly interesting as it included a number of specific examples and commentary on why a particular style is effective.  I also believe that I need to work harder at bringing life to what my characters are feeling in subtle but effective ways.

Having said all this, I think it’s important to keep one vital point in perspective.  There are many award-winning novels which are clearly labours of love by their authors, memorable for their readers, and which never make the bestseller list.