As most readers know, I recently published a nonfiction book titled How to Get Started as a Technical Writer. I published it only on Amazon, but in both a Kindle and print (Createspace) version. I also gambled by enrolling the book in Amazon’s KDP Select program.

If you’re a fiction writer–especially a genre fiction writer–the Internet’s rife with resources for selling your book. Nonfiction’s a different story; I find little real-world information about self-publishing nonfiction.

So, motivated by the shortage of real-world data, and inspired by David Gaughran’s blog about his publishing journey, I plan to share the first three months of my Amazon/KDP Select experience with you. Warning: This post may be deadly dull for a lot of readers, but I want to add my experience to the pool of self-publishing knowledge.

How I promoted the book

I did very little promotion, and didn’t “build up” to the publication date on my blog or other outlets. When I released the book, I wrote a blog post and Twitter message, then sent a similar e-mail to friends and mailing list members. I did that again the day before my one “free promo” day, then once more on the free day. Then, I sent out Twitter messages when the price dropped (three times). That’s it–nothing else.

KDP Select results

For March 2012, Amazon announced a–for lack of better term, I’ll call it a “dividend”–of $2.18 per Select borrow. That’s 15 cents more than February 2012 ($2.01). The total allocated fund was $600,000 for March, meaning there were around 275,000 borrows in March. I estimate there were about 298,000 in February–a notable drop in borrows. The fund for April 2012 is again $600K.

A grand total of three copies of my book were borrowed in the Select program in March, earning–I think–$6.54. Select program sales results are aggregated into the overall sales figures, so it’s hard to tell; I haven’t seen a full report yet for March.

Sales results

Thirty-one people paid good money for my book. Thank you!

No sales or free copies for France, Spain, or Italy. I guess my book didn’t translate well.

I sold at least one copy almost every day–a hopeful trend (and one that’s continuing in April).

One caveat: there was one Kindle return, but it’s unclear if it was a paid sale or Select borrow. It seemed to occur during a window when I lowered the price by $1, so I’m guessing someone saw the lower price, got a refund, then bought at the lower price. Now that‘s being thrifty. Also, the print version came out on March 19th, so it only had 12 days of exposure.

Pricing

I deliberately experimented with price on Kindle versions, starting off at $6.99. There were two sales. A week later, I dropped it to $4.99 (two more sales), then $3.99 for a week (four sales), and finally stopped at $2.99 (22 sales).

The “sweet spot” for my book’s price seemed to be $2.99. Prices of other books in my categories were all over the place, so I’m still not sure what the best price point is for my book.

Giving books away

I gave away 381 copies on my only Select free promo day. Based on traffic to my website, some Google search analysis and a bit of intuition, it’s clear most people found my book either from Twitter, or from sites that track and troll for free Kindle books. I found my book on a dozen lists of “free today”-style websites and Twitter feeds.

Did the giveaway have any effect on my paid sales? Not one bit. In fact, there were no sales for two days after the giveaway, and the sales trend never spiked afterward. I have four free promo days left; I may not use them.

Amazon rankings

Contrary to popular belief, Amazon’s ranking system isn’t a complete secret; it’s just difficult to understand and pin down with precision. Amazon uses a well-known system (that’s also now used by YouTube), but how they implement it remains partially hidden.

The algorithm’s net effect on sales? I’m not sure, though many people have tried to guess. One thing I do know–my book’s trip up and down the ranking continuum didn’t match any of those guesses at all, and other authors I’ve asked report the same thing–not even close.

The only data I have is one month of my own sales. Here’s what  I saw:

1. My book appeared in the “Also Bought” widget within hours of the first sale. It stayed there, too. I verified this using another account to surf Amazon.
   
2. Rank improved sharply an hour or two after each sale–and if two or three sales happened within hours of each other, even better. One sale was typically enough to bump the book from, say, 85,000 to 40,000, and into the top six books in the category.
3. Rank declined within a few hours after a sale, and  if several hours (say, four or more) went by without a sale, it would easily drop 100,000 or more. My book oscillated between the 30-40k range and the 80-140k range–but rarely in between for long.
4. Free books affected rank just like paid sales. This was surprising.
5. Rank had no apparent impact on sales. This goes against conventional wisdom, but I believe that most people searching for nonfiction e-books on Amazon find them in one of two ways: a custom search, or the “Also Bought” widget. Fiction may be different, but all of the top 25 books in my categories were displayed in the “Also Bought” widget. I checked.

Here’s where my book debuted, before any sales or reviews:

My book’s best ranking occurred on March 18th, two weeks later and a few hours after two sales:

During the “free promo day”, things looked very different:

And here’s the worst ranking, despite 20 sales and several five-star reviews:

Tracking tools

Amazon’s Kindle sales tracking tools are limited and somewhat clunky, both in appearance and functionality. With only slightly more effort, Amazon could offer much better (and more detailed) tools. Reporting and sales tracking is one area where publishers can eat Amazon’s lunch.

I created a NovelRank account, but it seemed unreliable. I often noticed sales appearing on Amazon’s site that never made it to NovelRank, and NovelRank reported slowly, often hours after the sale. It wasn’t critical to me, but I’d like to see the NovelRank’s creator be clearer that his tool provides “approximate” data. Will I keep using NovelRank? Probably not.

Summing it up

I didn’t know what to expect, but was pleasantly surprised that the book sold 31 copies (and hundreds of readers grabbed a free copy). I wish Amazon’s reporting tools were better. I’m not sure I’ll continue in the Select program for the full 90 days–as of today (April 10th), only four copies have been borrowed. It pays me about the same as a purchased copy, but at a huge opportunity cost–I can’t distribute the e-book anywhere else. Is it worth it? I’m almost ready to say no.

The books that routinely held top spots in my categories were high-profile books, written by “name” authors, with several peer blurbs and a history in print. My book hit the top five often, but never stayed there long. Worrying about ranking didn’t do me any good, because sales were unaffected–another surprise.

Stay tuned! I’ll return with more information about the second month sometime in May–and hopefully a lot more sales. Meanwhile, I’d be grateful to hear your thoughts on what you’ve read here. Did I miss something? Know of any strategic resources that might help? Please drop me a line, or share them in the comments below.