My attention was drawn to a set of numbers from BookBub available here: https://www.bookbub.com/partners/pricing?fbclid=IwAR1mMlf6nO5oAS5QN0pD-9V-NRAxkdNz6AQMNOJurGT1NnD_DQEjiPzOEL0
Some major caveats before we go into them. Firstly these are for marketing purposes and as they say “averages are based on historical data, but are only meant as a reference and are not guaranteed”. The book figures also only apply to free downloads and discounted book sales. Lastly, these are BookBubs numbers and other retailers of books may show different patterns.
A broader caveat to add when considering any kind of average sales within books (or other media) is the dreaded power-law distribution. A small number of books account for a large number of sales and conversely a large number of books account have small sales individually but account for a lot of sales together. The arithmetic mean has many flaws but it is particularly flawed in such circumstances. One huge hit (e.g. The Da Vinci Code) will have an outsized impact on the average book sales even if other books are selling poorly.
Tables and things after the fold…
The relevant figures provided (with variable symbols I’ve picked):
- BookBub subscribers by genre category [P]
- Free book average downloads [D]
- Discounted book average sold [S]
You can use these categories to compare the relative popularity of genres. For example Crime Fiction & Thrillers have the most BookBub subscribers but Cozy Mysteries & Crime Fiction have the most free downloads whereas Crime Fiction & Historical Fiction have the most sales.
Now, if you are genre shopping as an author, can you use this data to pick a sweet spot for selling books? For example, if Crime Fiction and Historical Fiction both sell well then maybe Historical Crime Fiction is the best genre? True Historical Mysteries do sell well but as they are a more narrow genre they sell less well then Crime or Historical fiction in general but there is a trade off there. A narrower genre definition must have smaller numbers but every book is a subgenre of 1. Broad genres sell more but they are also more crowded with books.
Last week our old chum Dave Freer posted about these numbers at Mad Genius Club (https://madgeniusclub.com/2020/02/17/super-readers-and-their-interests/). He was very interested in defining a sweet spot:
“Of course if the interest in that category is small, the download/sales numbers are small. Of particular interest is what translates into paid sales. The areas to look at warily are those where a lot of people will download for free, but few buy. The ideal ‘sweet spot’ is reasonable volume (high interest) reasonable percentage of free downloads, and high paid percentage. A high paid percentage may seem desirable but if that’s off low interest number, and low ‘exploratory’ (free download) percentage, that’s not too good for new authors, but possibly OK for established ones.”
That strikes me as an odd combination of factors. Given the numbers there is a simpler question that can be asked: which genre has the best chance of a book selling? I’ll come back to that.
Dave decides on a index that is calculated as:
“A QUICK ‘DIRTY’ CALCULATION OF NEWBIE SWEET SPOT ( a lot of interest – a lot of experimenting with new to them authors -free, and a high percentage of sales) RANKED (NUMBER INTERESTED*FREE PERCENTAGE*PAID PERCENTAGE)”
Using the variables I named earlier that is P*(D/P)*(S/P) or D*S/P. Consider for a moment what on Earth that might be calculating. It is simply arithmetic nonsense. In units it is downloads-squared divided by people. There’s no rhyme or reason to multiplying the free downloads by the paid-for downloads. I assume he thought he was finding a fraction of a fraction but it really doesn’t work out that way.
A rationale for this odd choice becomes clearer when Dave lists the top 16 fiction (maybe) sub-genres. By his index the LGBT and African American Interest end up at the bottom of his list.
A sweet spot is easier to find within the limitation of the numbers we are given. Firstly we can put the subscribers number [P] aside. That number will largely be determined by how we define categories (broad categories have a bigger P). We want to make an estimate of how well a book might sell (it will be flawed and probably not a very good estimate). We don’t need the P number because we have actual sales. We do want to factor in how crowded the subgenre might be and we can work that out using Free Downloads plus Sales [D+S]. We can add these numbers because they are the same kind of things (i.e. essentially the same units).
Sales divided by total downloads (free and paid) gives a meaningful (up to a point) index [S/(D+S)]. It is the proportion of sales compared to the total number of downloads. That gives you Table 1 below. LGBT as a category does very well. Why? Because even though it is a small category, more of the books in the category SELL. Christian Fiction does not do well here because it’s proportionally a lot more free downloads.
If you want to throw the subscriber number in you can but what you end up with is what I’d call bookshop categories (Table 2). Literary fiction tops the list because it is a very broad category and it has proportionally a lot of sales. Table 2 is very unremarkable.
|Politics and Current Events||20.00%|
|Religion and Spirituality||10.77%|
|African American Interest||9.09%|
|Biographies and Memoirs||8.51%|
|Dark Romance & Erotica||6.89%|
|Time Travel Romance||6.62%|
|Advice and How-To||6.34%|
|American Historical Romance||5.69%|
|New Adult Romance||5.56%|
|Action and Adventure||5.08%|
|Teen and Young Adult||4.88%|
|Biographies and Memoirs||280,738|
|Religion and Spirituality||132,462|
|Action and Adventure||131,578|
|Politics and Current Events||98,000|
|Advice and How-To||93,263|
|Teen and Young Adult||81,009|
|New Adult Romance||74,489|
|American Historical Romance||68,842|
|Dark Romance & Erotica||45,451|
|African American Interest||41,818|
|Time Travel Romance||39,054|