epub2 missing

It seems epub 2 is missing from Compile For.

I use epub 2 because of the way Amazon generates KENPC. Using epub3 means I have a lower KENPC, thus get paid less per borrow.

ePub 2 is no longer, and there is no bringing it back. If you really need it, a conversion tool like Calibre would be the way to go.

But wouldn’t using Mobi be the best choice for uploading to Amazon anyway?

That is incredibly disappointing. Yes, I upload a mobi, but the mobi Scrivener creates is inferior in terms of KENPC. I compile to epub2, edit it with Sigil, then use an Amazon tool to convert to mobi before uploading. Directly uploading the epub also results in lower KENPC.

It ranges, but using epub3 tends to mean a difference of about -5%. That’s thousands of dollars lost.

As for Calibre, I don’t see a way to do 3->2.

Why was epub2 removed?

Mainly because the conversion process used to generate it was subpar, and very difficult to make any improvements to. We’ve been unable to provide internal cross-references and other rudimentary capabilities for years, because of that. Secondarily the new format will be much cleaner and easier to design with as a result (no serialised class names in HTML elements for example). But it’s also time, as well—ePub 2 hasn’t been a required format for devices for a long while now.

I don’t really understand how a difference in format would cause a difference in “KENPC” (I had to look up what that even means). If that is as I understand it, an attempt to quantify eBook material in terms of physical sheets of paper in a bound book, wouldn’t then design and layout be of more impact than the internal specifications of the HTML and XML files?

I know nothing about this, obviously, but I’m trying to think of how the underlying problem can be solved.

KENPC is part of how we get paid. A number is assigned to each book by Amazon, and authors get about .0044 dollars per page read. So for a book with 400 kenpc, if someone read the whole thing, the author would earn $1.76.

No one really knows how KENPC is generated. Amazon tells us almost nothing about how it’s generated. Formatting used to change it, but now not so much. I have no idea why I get more KENPC with epub2.

That might be something worth checking with them on. I can see why they may not want to reveal the whole algorithm, to avoid people gaming it, but they should at least have a response for why two different specifications yield different results.

I suspect Amazon will claim (accurately or not) that .mobi gives more accurate results than epub2, which is one of the reasons why epub2 is being deprecated.

But yes, this is a question for Amazon.


Open Calibre.
Select the doc to be converted.
Convert, Convert Individually (bulk convert will work too, with one change)
Epub output (nearly the last item down the left side of the page)(if you use bulk convert, this is the last item down the left side)
7th line down in the main text on that same page: Epub Version. Enter a 2 (it defaults to this, btw, and the text below suggests not using 3 unless you need it).

Should just work; I tried it with Scriv’s Epub 3 output, and it seems to be fine (I’m not really familiar with Epub details, tho, so your mileage may vary).