Friday, April 26, 2013

Some Staggering Statistics on Self-Publishing

The best estimate of Self-Publishing titles I've found is 1.2 MILLION in 2011.

During the various debates on the rightness or otherwise of the market, various people have pointed to certain successes.

You might name half a dozen big hitters.

You might research and come up with 200 authors doing very well.

You might REALLY do the digging and come up with 2,000 Names.

I was driving to the station this morning, and having once done a course in statistics I paused a moment and thought, 1,200,000 titles (2011) and growing.

Shouldn't I EXPECT the occasional freak result?

Now look at this

1,200,000 Titles 90.00% bad = 120,000 decent-to-best-sellers

1,200,000 Titles 99.00% bad = 12,000 decent-to-best-sellers

1,200,000 Titles 99.90% bad = 1,200 decent-to-best-sellers

1,200,000 Titles 99.99% bad = 120 decent-to-best-sellers

Is there anybody arguing that there are more than 12,000 
(TWELVE THOUSAND) SP titles doing well

If the figure is 12,000 that is just ONE per cent of SP

But are there 12,000 titles doing really well? 

You tell me. 

Even if you got to 120,000 that would still mean that the market you speak of is 90% junk.

There is an excellent academic book by Nisbett & Ross called (roughly) Errors and Strategies in Human Inference 

It looks at all the inbuilt statistical fallacies we carry inside us like thinking red is "due" at roulette because black has come up five times. 

Here is one. You are about to study for a degree and you really need to get a "First" 

There are two universities. One only gives a first about once every ten years and it gave one two years ago. The second give 2-3 "Firsts" a year. 

Now, when asked which university, virtually all respondents pick the second university. 

That's pretty much a no-brainer. 

But here's the oddity. If you now say to the respondent that they KNOW the person who got that "First" at the almost-impossible university, the choice of university switches from 99.9% Uni 2, to an almost 50-50 split. 

That is people use the statistics correctly UNLESS they believe they have inside knowledge. That inside knowledge completely distorts their choice 

And of course, that's what is happening with the occasional WOW HE MADE IT stories in self-publishing. Tell me one, and I might ignore the stats, tell me ten and now I think success is likely, tell me of 100 and I think I'm going to be rich. 

And yet 120 is just 0.001 of a chance, one in a THOUSAND. 


No comments: