An interesting article was posted on Monday about the relative sizes of the various ebookstores. It took me a couple days to think about and double check his methodology. I'm now confident that I can tell you the data is not valid and I can show you why.
The author didn't go far enough. I found the error in the data because I checked more stores than he did. He only checked 4 ebookstores, and among the ones he skipped were Borders, Kobo, and Smashwords. There are others, yes, but most all of the smaller ebookstores are honest about the number of titles they sell so there is little reason to double check.
I'm going to take you through the methodology and and the results, then I'll explain why the results are invalid.
First, a word about Smashwords. This was the one honest site of the bunch. Their about page says 18k titles and I found 22k. Obviously they've grown since the last time that page was updated.
I used Google to search the ebookstore sites. In this day and age, a professionally made site should be set up so Google can find every page. (If it's not, then it's the fault of the developers, not me.) The Google searches fell into 2 types. Some sites (Kobo, Sony, Smashwords) had their ebooks organized in a particular location. Other sites required me to identify and search for a particular term found on each ebook listing. Here are the searches, if you're interested:
- Sony (www.google.com/search?q=site:ebookstore.sony.com/ebook)
- Kobo (www.google.com/search?q=site:kobobooks.com/ebook/)
- Borders (www.google.com/search?q=site:borders.com "About the eBook")
- Smashwords (www.google.com/search?q=site:smashwords.com/books/)
- iBooks (www.google.com/search?q=site:itunes.apple.com/us/book "This book is available for download with iBooks")
- B&N (www.google.com/search?q="buy this ebook" site:search.barnesandnoble.com)
You might have noticed that I left Amazon out of the mix. If you ran a similar search for them you'd get around 700k titles.That's about the number they claim to have (even though most of the titles are in the public domain).
- Sony (57k)
- Kobo (128k)
- Borders (92k)
- Smashwords (22k)
- iBooks (33k)
- B&N (29k)
Smashwords were the key to finding out the results were invalid. You see, Smashwords are both an bookstore and a distributor. Their titles are available from Sony, B&N, and iBooks.
Think about it. If Smashwords are providing 20 thousand titles to B&N then all of B&N's other distributors account for only 9k. Does that make any sense? The same applies for iBooks. Do you really think that all of the Agency 5 only account for 13k titles?
It was an interesting idea, but unfortunately this method doesn't work.