Originally Posted by Dahak
That data is absolutely wonderful. Thank you for tabulating it for us.
But there seems to be a bit of a discontinuity in the summer of '08.
It's been a while since I've played with numbers like this, but I believe that if you fit the curve from the start to about June or the beginning of July '08 and do a separate curve fit for Sept '08 onward, both curves would a be quite a bit more linear.
Yes, there is some acceleration, but that extrapolated end-point would be lower if you start the curve-fit in Sept '08, instead of Nov '07.
Thanks again for the numbers.
I've played around with it to know what you're talking about, but the trend is still accelerating. Another way to test the validity of the curve is to see what it looked like with earlier data, and the latest data points are coming in on target. The projection isn't being "flattened" over time to fit the data.
However, having done this sort of charting before in other industries, it's really impossible to tell the difference between a linear trend and a polynomial one when you're at the low-end of the curve; to the extent that there are errors and other external factors in the counting you really can't be sure which is which. Right now it appears that a year ago the average was more like 300 books a day and today it's about 500 books a day, but it could turn out over the next couple of years to be a linear trend that averages 400 books a day. You just never know for sure.
That said, publishers tend to produce a set amount of new titles a year. As more publishers embrace ebooks, and as more backlist titles are converted to ebooks along with new titles, it only makes logical sense that the rate at which ebooks become available would be increasing, not constant. (That's assuming Amazon is successful as a retailer of ebooks, of course. If you are anti-Amazon and think most publishers are going other routes for selling their ebooks, then naturally you might believe these numbers must be flawed.)