[Michael_J_Sullivan]

I keep hearing how the industry standard for ebooks is running around 25% - 30%, which shows no correlation to my own numbers. Certainly in the self-publishing arena ebooks dominate, but as my own traditional numbers run more along the lines of: 25% audio, 45% ebook and 30% print I just don't have a good feel for what is real and what is speculation.

To try and help, I've made a survey and would like as many kbers to help provide their numbers (especially since so many surveys like the recent one from DBW didn't get this group of writers information). It doesn't require you to expose the number of sales or your income, just the percentages between the various formats. In the spirit of Hugh Howey's work on author earnings, the data is freely available to everyone. You can post anonymously if you prefer by not using your real name, but as it is just %'s and not income I hope that most will use their own name - you can see a full breakdown of all my traditional results to date in the data now.

Thanks a ead of time for anyone who contributes, trying to get "real data" in this crazy world of publishing hasn't been easy.

 

[Ronny K]

Good idea, Michael... I'm curious to see the numbers.

Question: how relevant is the print percentage to a group of authors who mostly don't get brick&mortar distribution?

 

[zoe tate]

Question: how relevant is the print percentage to a group of authors who mostly don't get brick&mortar distribution?

This is the important question. A survey completed by a self-selected group of mostly self-published authors who mostly have no bricks-and-mortar distribution is going to produce results relevant almost exclusively to a self-selected group of mostly self-published authors who mostly have no bricks-and-mortar distribution. I don't suggest that that in any way devalues its potential findings, of course, but it obviously limits their relevance enormously.

 

[Michael_J_Sullivan]

Good idea, Michael... I'm curious to see the numbers.

Question: how relevant is the print percentage to a group of authors who mostly don't get brick&mortar distribution?

I think it is important if they are trying to decide whether to go hybrid. If print is 70% then doing at least one book traditional may make sense...if it's 10% not so much.

 

[Michael_J_Sullivan]

Oh, yes! Fantastic idea.

I keep hearing that 'print is 70%', and I just don't believe it. In fact, I tune right out when anybody says (writes) it.
I just asked DWS in the comments of his blog what his percentages are since he is so pro-print. He was kind enough to reply. http://www.deanwesleysmith.com/?p=11835#comments
Not 70%, but good enough to make me look at my print and think I am doing something wrong. Mine were 76% ebook/23% audio/1% print last month (which is a pretty low ebook month and my first audio month). I finally raised my print prices to DWS's suggested $2 royalty for extended distribution and have sold more books in the last 10 days than in any other month and have made more money than all of last year for print (it wasn't a big number to beat)...so maybe those percentages will be a little different next month.

And I will go fill out that survey.

Thanks for adding our data to the pool!

 

[Michael_J_Sullivan]

This is the important question. A survey completed by a self-selected group of mostly self-published authors who mostly have no bricks-and-mortar distribution is going to produce results relevant almost exclusively to a self-selected group of mostly self-published authors who mostly have no bricks-and-mortar distribution. I don't suggest that that in any way devalues its potential findings, of course, but it obviously limits their relevance enormously.

The plan is to have authors of all types do the survey: self-published, hybrid, traditional. I posted it here, but I'm also going to reach out to those areas where more traditional published authors run. There is a field to indicate how the book was published, so you can filter the data for only those that were traditional if you are interested in going hybrid. It may turn out that when filtering for self - the sales are higher than we think. Bottom line, we won't know what to make of the data until we have enough to analyze...and since the data is available to all, people can do their own analysis to see how relevant it is.

 

[PhoenixS]

***********

 

[Michael_J_Sullivan]

I don't even know how our sales could be caveated.

We don't have print rights to 37 of our titles, which also means they can't be included in any print versions of our single-author boxes. We have maybe 16 of those single-author boxes, so that's 53 titles we can't do print for regardless.

We've had 140,000+ sales of short-term multi-author sets - but print isn't even a market consideration for those boxes. It's apples/oranges.

Then, since our ebook inventory is mainly backlist, if we look at the print that HAS sold, it's around 60 million copies (very conservatively). If we use those lifetime numbers, it's like 98.5% print, 0.5% audio, and 1.5% ebook.

Yeah in your case it doesn't make much sense. But hopefully there will be others for which it does.

 

[zoe tate]

The plan is to have authors of all types do the survey: self-published, hybrid, traditional.

Even so, Michael, the statistics it produces will have broader validity only if the proportion of those "types of authors" exactly mirrors the proportion of each that exists with regard to their overall sales. In other words, you'd need to know to start with exactly what you're trying to discover, for the findings to be valid. No criticism implied, but it's a classic example of the famous "statistically circular argument".

The easy mistake is to imagine that if you get enough input into it, i.e. if the sample size is "big enough", its validity will increase. Unfortunately, this isn't so. What will actually happen, as the sample size grows, is that you'll simply be dealing with a larger self-selected group rather than a smaller one. The results will still be distorted if you try to apply them to "the general situation".

It isn't your fault: it's just "how it is". Hence all the mistaken conclusions that people so commonly purport to draw from trials of this nature. If the inputs are all self-selected, the output is, by definition, distorted.

Forgive me if I sound critical, please: you probably weren't expecting a statistician to comment at all! :-[

 

[Michael_J_Sullivan]

Even so, Michael, the statistics it produces will have broader validity only if the proportion of those "types of authors" exactly mirrors the proportion of each that exists with regard to their overall sales. In other words, you'd need to know to start with exactly what you're trying to discover, for the findings to be valid. No criticism implied, but it's a classic example of the famous "statistically circular argument".

The easy mistake is to imagine that if you get enough input into it, i.e. if the sample size is "big enough", its validity will increase. Unfortunately, this isn't so. What will actually happen, as the sample size grows, is that you'll simply be dealing with a larger self-selected group rather than a smaller one. The results will still be distorted if you try to apply them to "the general situation".

It isn't your fault: it's just "how it is". Hence all the mistaken conclusions that people so commonly purport to draw from trials of this nature. If the inputs are all self-selected, the output is, by definition, distorted.

Forgive me if I sound critical, please: you probably weren't expecting a statistician to comment at all! :-[

I agree, there are inherent problems wit "self-selecting" and people should keep that in mind when looking at them. Still, there is so much secrecy in publishing at all levels - sharing contract language, sales data, that I think it is worth exploring. Maybe it will go no where and do nothing. But I wanted to at least give it a try.

 

[PhoenixS]

***********