Kobo Report on Publishing in the Era of Big Data

[Ted]

An opportunity lies among those books that have high completion rates yet suffer low sales. Clearly those readers who have stumbled upon these books have loved them – so while the marketing team or editorial department may not have seen a winner among these titles, the reader did; it may well be worth experimenting with marketing efforts to see if the books catch on with a wider audience once they’ve been surfaced and attention is drawn to them.

Over all, being able to identify true reader engagement allows for more targeted allotment of marketing and publicity resources. For single titles that are over-performing in terms of engagement versus sales, is there potential to create more awareness, a follow-up title, or turn the title into a series of books?

This report from Kobo looks at data in the same manner as Mark has done with his reports, but perhaps making the use of such data more clear to some authors. Data from subscription services like Scribd and/or Oyster provide valuable information on completion ratios of your ebooks by readers, thus providing more insight for authors of which works are consistently attracting readers and which works are not.

Kobo big data report in PDF Publishing in the Era of Big Data - Kobo Wh....pdf (851.82 KB)

[serbanvcenache]

My books still have yet to show up at Kobo or Flipkart.

[Ted]

Oct 14, 2014 9:27:50 GMT -8 serbanvcenache said:

My books still have yet to show up at Kobo or Flipkart.

One has... store.kobobooks.com/en-US/ebook/a-heretical-divide

[serbanvcenache]

LoL. That eluded me; maybe next year the other two will appear as well.

[Ted]

Tell Smashwords Support your ebooks haven't yet appeared and it reflects badly on Smashwords. It is up to retailers to display a work or not, but it doesn't hurt to keep Smashwords informed of the situation until it is corrected.

[chelsfield]

Ted, just got finished reading this report. It makes for instructive and fascinating reading. Fascinating, to my mind, because there is now a whole set of data not provided by print books: engagement rates, and more specifically completion rates. That we can now know how many readers actually finish our books is great news (even though it still can't tell us exactly why). I guess as a reader it feels a bit on the creepy side....

In a sense, publishers have no excuse to go only by the bottom line, as the data shows that there is a much more nuanced way of understanding what defines a successful title. Let' s see if it makes a difference. 

Would that we could get that same kind of statistical feedback from Smashwords (yes, yes I know why we can't).

Thanks for a good read. 

[Ted]

Oct 28, 2014 1:24:31 GMT -8 chelsfield said:

Ted, just got finished reading this report. It makes for instructive and fascinating reading. Fascinating, to my mind, because there is now a whole set of data not provided by print books: engagement rates, and more specifically completion rates. That we can now know how many readers actually finish our books is great news (even though it still can't tell us exactly why). I guess as a reader it feels a bit on the creepy side....

In a sense, publishers have no excuse to go only by the bottom line, as the data shows that there is a much more nuanced way of understanding what defines a successful title. Let' s see if it makes a difference. 

Would that we could get that same kind of statistical feedback from Smashwords (yes, yes I know why we can't).

Thanks for a good read. 

Smashwords does provide completion rates through Scribd, which is not hugely helpful in the aggregate but does allow authors to have some data on completion rates of their works. My Scribd data isn't deep enough - less than 2 years of data is quite useless - so when one of my ebooks has completion rates of 52, 99, 100, 39, 47, 94, 82, and so on up to about 100 different rates the data is pretty useless.
 
It just tells me some people liked it and some didn't like it enough to read it all. Now, I must be fair in mentioning all my ebooks for children have many pages in the Appendix which are images of characters in the story and are made to be printed out. So if I have 20 pages for coloring out of a total of 60 pages then completion rates for those ebooks are changed as story ends about 2/3's through ebook.

What would help would be to know if reader was female or male, age of reader, time of day read, what other ebooks reader purchased or read, how often reader returned to read book and where he/she started and stopped reading and time of day reading started or stopped. All this data is available to ebook retailers but not to authors.

I agree it is a bit creepy the amount of data ebook retailers collect on a readers habits. But that's today and the way society works. Here's a link to a TED talk on why privacy matters in this day and age.

[chelsfield]

So perhaps your Scribd numbers are generally communicating that readers are finishing the stories but not availing themselves of the coloring pages? If that is a fair assumption, would you consider not including those pages and just expanding the story? Of coorse, you could try an experiment and produce one title both ways...

[Ted]

The picture ebooks for children are selling well, and your excellent point about expanding the stories is something I've been considering. I've been thinking of doing a sequel for some of them, so the coloring pictures in the Appendix wouldn't have to be in one version of a sequel. Great idea. Thanks Chelsfield.

[chelsfield]

Would make an interesting experiment. Good luck!

[Ria Stone]

Ted, you are a wealth of information. While I read through the report, I think I need time to digest it because it approaches "reading" from a different angle. But, very interesting. Thanks.