If you want to estimate the sales of your own book, one of the best ways is to estimate the sales of the closest competition. There are other sources of sales data, such as Ingram’s iPage, and Bookscan, but I’m going to discuss the more complex process of turning Amazon sales ranks into estimates of sales through Amazon here.
Their sales rank formula shifts and is proprietary, of course, but one thing we do know is that it’s always roughly inversely proportional to sales, over any reasonable range. Most of you just came to a screaming halt, didn’t you? Well, it’s not as arcane as that sounded. You can do this, really. Yes, really.
Once more, in English, and taking one step at a time:
If you look at books that are selling roughly similar amounts, then you can say that Amazon’s sales ranks can be inverted (divide 1 by the rank), and the result will more or less proportional to sales.
Proportional to the sales means you can find a number to multiply the inverted ranks by, and you’ll get an estimate of sales through Amazon.
To find the number, I recommend that you look at Morris Rosenthal’s Foner Books site. He has a page, here, that gives a graph of sales per week and sales rank. I’ve never found a better correlation that’s publicly available.
Pick a sales rank number that’s in the middle of the range you expect your book to inhabit. Go down that line in the graph until you hit the red line, and then run across to the sales per week axis. Jot that number down beside the rank. (Oh, and yes, the graph looks a little different than you may be used to, but don’t worry, it’s more accurate because the axes are logarithmic.)
Now, multiply the sales per week (or year, if you prefer) by the rank. That’s the factor of proportionality. If you divide your resulting number by sales ranks that are close to the one you chose, you’ll end up with an estimate of the sales for that title.
Caveats: First, sales ranks change quickly. Use the middle of the range that each title inhabits.
Second, you’re only estimating sales through Amazon. Only you know whether that’s likely to be 12% or 75% of the sales for this title.
Third, this is most useful for estimating the sales of titles that are competing with yours, or for estimating those numbers in order to get a rough ballpark estimate of the potential sales of your future title.
And last, but never least, if any of the calculations I describe seem to you to be giving you a wrong answer, if the answer is just plain “off,” rely upon your experience and judgment, not on the spreadsheets.
So, who has a question about this? A suggestion for improving it? Or something else they’d like to raise?
Ive been studying that chart on foner books for a while now. As Im trying to come up with the right formula to convert the sales rank into a books sold per week number.
Now I understand its logarithmic… but the chart dosnt make sense to me. Because the lines are uneven on both axis of the chart.
Each range is divided by 10, thats for both axis. So do you take half way to be line 5? Or do you take half way to be half the physical measure along the axis?
The reason I ask (Im no mathematician) is because if you use the uneven lines to find the related number on the other axis… theres no point in having it logarithmic at all. It defeats the point. The chart in order to be correct should be even on one axis.
My guess has always been, that it must be and that the lines are supposed to help you estimate the number your after… but, which axis is even, which is not.
The chart is not very well explained.
Thank you for commenting, David. I overlooked the fact that not everyone knows how to use a logarithmic scale. I’m going to do a post on that issue, right now.
[...] commenter on Estimating Sales, Part III was having difficulty with the logarithmic nature of that graph. So, rather than answer the [...]