When the pagecount-based Kindle Unlimited 2.0 compensation method was first announced, we were initially worried about our ability to model it accurately in our Author Earnings reports.

But it actually turned out to be less of a problem than anticipated, because when Amazon announced the total number of Kindle Equivalent Normalized Pages (KENP) read in July 2015, the July 2015 KU $ “pot size”, and the $0.00578-per-KENP payout rate, they effectively handed us an easy way of checking our model’s accuracy.

Traditionally-published authors and Amazon-published authors get the full sale price for each KU download. So for AE purposes, we just needed to calculate KENP payouts for indies.

We knew that our September 2015 spider run collected the details of **60.9%** of Amazon’s daily paid downloads (direct sales+KU borrows), of which a total of **245,000** were of indie books in KU. We knew that our data set represented **60.9%** of the total, because we knew which ranks our spider captured the sales from, and which it didn’t. We treat the other **39.1%** it didn’t capture as being distributed similarly — which is the maximum-likelihood hypothesis, absent any evidence or compelling argument to the contrary.

For September, this meant that there were:

**402,300** daily paid downloads of indie books in KU

Some were direct sales, some were KENP-compensated downloads. For our charts to be accurate, we had to separately calculate the author earnings from each.

We already knew the regular page counts of each of the books in our data set (it’s one of the pieces of data grabbed by the spider — shown for all books in column **AD** of our September 2015 spreadsheet). So the only remaining parameters we needed for our model were:

— **KUBVB%**: the Average borrow-vs-buy percentage for KU indies

— **KENPIF**: Average pagecount-to-KENP inflation factor

— **KURT%**: Average KU read-through percentage for a borrowed indie book

For **KUBVB%**, a Kboards poll had given us a starting point of * 50% borrows/50% buys*, which we later validated to within a few percent against Amazon’s KU 1.0 payout and pot size announcement.

The switch to KU 2.0 changed * author compensation*, obviously, but there’s no real reason for it to change

*. So we stuck with 50/50.*

**reader behavior**For **KENPIF**, I’ve seen it vary widely (my own books average 1.9), but most indies are reporting an KU-to-KENP ratio between 1.5 and 2.0. I plugged in 1.6.

For **KURT%**, taken on its own all we can say for sure is that it’ll lie somewhere between 0% and 100%… which is hardly helpful.

But here’s the kicker:

* We also know from Amazon’s July KU 2.0 announcement that the total Kindle pot size was $11,500,000 and that the per KENP payout was $0.00578*, so we also know the following

**must**be true:

(**KU indie paid downloads** * **regular page count of each**) x **KUBVB%** x **KENPIF** x **KURT%** x **$0.00578** = **$11,500,000**

So plug in any reasonable estimate for **KENPIF**, and the above the formula gives you the corresponding **KURT%**.

And even better, any combination of **KENPIF** x **KURT%** that gets you to the **$11,500,000** payout will give us the correct indie KU author earnings, whether it’s a **KENPIF of 2.0 multiplied by a KURT% of 68%**, or a **KENPIF of 1.43 multiplied by a KURT% of 95%**, or **anything in between**. For the spreadsheet, we went with a **KENPIF of 1.6, and a KURT% of 85%**, which hit the $11,500,000 right on the nose. But if you don’t like those numbers, feel free to plug in your own into the spreadsheet and let it recalculate. As long as the combination results in $11,500,000 of KU page-based payouts, it won’t change author earnings… or anything else, for that matter.

If you stuck with me through all of that, congrats. You’re a fellow data geek. 🙂

Wow, thanks for posting these details! It all makes sense. I don’t think there’s any way KURT is 85% but as you noted, it doesn’t really matter. The total pool is the same. The two numbers that could throw things off are the KUBVB and the sales-rank ratio. My guess is that when you update the sales to rank ratio as you mentioned in the main thread you would soon, KURT comes down.

Awesome explanation. Thanks!

Math … numbers … woah, paycheck? Great!

It’s very tidy that the unknown and the shakiest estimate can be treated as a single factor. Of course KUBVB% is also an estimate, but you could validate it against the KU 1.0 reported numbers.

If KENPIF could be firmed up by sampling more titles, KURT% would also be firmed up by proxy. We know what their product has to be.

It would be interesting to compare the KURT% to the sales conversion rate of samples downloaded. I know in my case the conversion rate is less than 25%.

Thanks for all your hard work.

Very interesting article indeed. However, I’d like to point out from the experience of my own books that KUBVB% varies significantly based on the price. It seems that KU users feel that they get a better bargain if they load a higher-priced book. The number of loans (when compared to actual book pages) for my more expensive books is significantly higher than for the cheaper ones.

However, this detail might be skewed by the fact that most of my books are in Spanish, and the Spanish-language market might differ from the English one.

It’s very tidy that the unknown and the shakiest estimate can be treated as an unmarried element. Of direction, KUBVB% is also an estimate, but you could validate it in opposition to the KU 1.0 pronounced numbers. The Very thrilling article certainly. however, I’d want to point out from the experience of my very own books that KUBVB% varies drastically based totally on the rate.

Thank you already publish this article, it is very useful for us.

what is mesothelioma and how did i get it

At the point when the page check based Kindle Unlimited 2.0 remuneration technique was first declared, we were at first stressed over our capacity to display it precisely in our Author Earnings reports.

since we knew which positions our insect caught the deals from, and which it didn’t. We treat the other 39.1% it didn’t catch as being disseminated comparably — which is the greatest probability theory, truant any proof or convincing contention actually.

According to Author Earnings, less than 45% of those author-earnings are going to traditionally published authors.IT Experts Agency has rapidly recognized and implemented IT solutions with an incredible reputation of getting results for our clients.