Where ebook buyers live (in the US)

I have a fairly deep backlog (50+ items) to blog about, so today's post will come from that archive of interesting tidbits. A couple weeks ago there was an interesting posting on TeleRead on where e-book buyers live in the US. I thought this would make a nice light topic for a weekend post.

While the article points out the limitations to its method, as a baseline set of numbers it is somewhat interesting. Not surprisingly, the states with larger populations account for the largest number of ebook sales, with five states accounting for nearly a third of all ebook sales.

Top five states by percentage of total ebook sales: 1. TX 8.57%, 2. CA 7.99%, 3. NY 5.99%, 4. FL 5.93%, 5. PA 4.13%.

Bottom five states by percentage of total ebook sales: 46. ME 0.32%, 47. SD 0.29%, 48. DE 0.29%, 49. WY 0.26%, 50. VT 0.20%.

The District of Columbia ranked last at: 0.09%.

What is interesting is what happens to the rankings when the numbers are adjusted to a per-capita basis. The ratio tells us the propensity to buy e-books in those states. A ratio of 1.00 would be average. A person living in Alaska, with a ratio of 2.92, is nearly three times more likely to buy an ebook than a person in an average state. In contrast, someone in California (which has the second largest percentage of ebook sales) is about two-thirds as likely to buy an ebook as someone in an average state, and has the lowest e-reader rate per captia of any state.

Top five states by e-book per capita ratio: 1. AK 2.92, 2. ND 2.29, 3. UT 1.58, 4. WY 1.44, 5. VA 1.41.

Bottom five states by e-book per capita ratio: 46. HI 0.77, 47. WV 0.76, 48. ME 0.75, 49. MS 0.75, 50. CA 0.66.

Interestingly, as with ebook sales, DC falls well below the average with a per capita ratio of 0.46.

So what is happening in the U.S. Capitol? Would the stats look similar among other US cities? One theory why Alaska and North Dakota top the list might be the distance between cities or access to local bookstores. Anyway, it is some interesting analysis and as the original article suggests, it might be worth applying the original data set against other factors to see if there are other meaningful conclusions to draw. Perhaps that is a post for another weekend. :)