A quick test recently confirmed that when a customer borrows your ebook and reads it through the Kindle Unlimited program, that counts as a sale the same way as if they had purchased your Kindle book directly. To demonstrate this here are is the salesrank data for the test book:
This is just a quick post that helps out the majority of users of NovelRank. I was able to request an increase to my API calls/hour with Amazon.com, meaning that within 1 hour I can make 10,800 API calls rather than 3,600 API calls. This means that (especially for books that have sold a copy in the last few days), sales rank data is being collected 100% faster than 3 months ago (~19/day vs ~9/day). That means that sales estimates will be more accurate moving forward (still underestimated for bestsellers) since the sales rank data it is based on will be more reliably captured.
One final note, some are turned off by the 100 book limit for user accounts, even for PRO accounts. The above is the reason this exists, because the more books in the system by a few individuals, the larger the slowdown they can have for everyone’s experience. NovelRank has always been about helping out the little guy authors, and that won’t change even if it means leaving money on the table.
As a reminder, PRO customers always get sales rank checks 23 of the 24 hours each day, among many other benefits.
Image credit: CC-A Marcelino Rapayla Jr.
This is just a quick update to let you know that the background checks that power the hourly data collection for NovelRank have been completely rewritten. While they have had improvements and tweaks over the last 6 years, this new design is leaps and bounds better since being rewritten.
- Each domain runs its own independent script. Previously, only Amazon.com and Amazon.de were independent. This means no more books on say, Amazon.fr, that are active but don’t have a new sales rank for days or weeks due to script inefficiency.
- More books checked per hour. When NovelRank started, Amazon restricted API requests to a certain number per hour. Thus, NovelRank’s scripts were designed around that. Now, they are throttled based on API calls/second. Thus, the new scripts can slow down when Amazon asks, but otherwise make better use of every second in every hour to get sales rank data.
- PRO user books every hour, on the hour. Previously, NovelRank put PRO books first as promised, however it would then run the rest of the books which sometimes took more than 1 hour to complete. It would not check PRO books again until that was done, so PRO books would be first, but still may miss an hour of sales rank data. Now, as soon as it is a new hour, PRO books are processed, pausing any current processing, ensuring hourly sales rank data at a consistent time of the hour, each hour.
In the continued effort to keep things running smoothly for everyone, it’s time to make-good on the agreement outlined in the FAQ over 5 years ago. Up until now, no old salesrank data has ever been removed. That means, if you started tracking your book 5 years ago, there are 30,000 salesrank records in the database for your book!
During the month of September, that number will be reduced to the last 2 years worth of data only. This doesn’t include estimated sales numbers, just salesrank records. If you would like to download them prior to them being expunged forever, you should go to a book’s detail page and click the Download to Excel/CSV option to show you the appropriate download link.
In the rare situation that a book is in the system, and after removing all but the last 2 years worth of salesrank data, it doesn’t have any new salesrank information, it will be completely removed from NovelRank.
I’m sure most people really won’t care about this, as the historical salesrank information is not exactly captivating, but I did want to give everyone a heads-up.