When I tell people my profession, everybody says the same thing: they want to write a book. But they won’t bother with publishers. They’ll do it direct. Self publishing. Print on demand.

I’ve done that. I have a tale to tell.

For those who don’t know me, my name is Keir Thomas. I’m the author of many books about Linux released through major publishers like Apress and Pragmatic Bookshelf, starting with Beginning SUSE Linux in 2004. I’m a full-time writer but I’ve also worked as a computer magazine editor and book editor.

I had to do it

At the end of 2008 I got an idea. It was the kind of idea you have no choice but to follow through. How about a pocket guide to Ubuntu, then the hottest distro around? You know, like the books by a certain publisher whose name sounds like Oh, Really?

I touted the idea to a couple of publishers (including Oh, Really). Nobody was interested. The profit margin is too low on cheap books, they said. I didn’t blame them: The economy was beginning to nosedive.

So I decided to publish it myself. It took me around three months to write and get everything together, but eventually I had a finished book and a website. I acted as my own editor (with the help of a few others to tech review), and even designed the cover of the book.

Back then everything was sliding into place for a self-publisher with ambition. CreateSpace, a print-on-demand (POD) off-shoot of Amazon, offered a pretty good deal.

Give it away now

I knew that publicizing the book would be difficult. If you’ve ever tried to get people interested in something you’ve made, you’ll know how hard this is: Why should anybody care?

I hit upon an idea: Why not give away the eBook (PDF) version? I could use Amazon S3 for hosting the file, so it would cost me just a few dollars per month.

Sure enough, giving the eBook away generated a lot of publicity. It was featured on high-profile blogs, including Lifehacker. Every now and again somebody new will blog it about, and I’ll see a spike in my website traffic.

I originally intended to price the book at $9.95, to compete with similar O’Reilly books, but $12.99 became a more realistic price (I eventually did experiments and found the book sold better at $12.99—consumers can be odd like that).

I get $4.95 in royalties for each print copy sold via Amazon and $2.35 for each copy sold wholesale (that is, to bookshops). Both figures are far better than the 10% royalty rate usually offered by publishers.

A big issue was that, for the first year the book was on sale, it was only available via Amazon.com. That was the only retail channel CreateSpace offered. I believe Amazon owns around half the retail book market in the U.S. but this was still a limiting factor. Additionally, overseas Amazon sites didn’t (and still don’t) retail CreateSpace books.

The big question is: As I write this, am I reclining on a chaise longue stuffed with $100 dollar bills? Did my self-publishing experiment work?

The short answer: Not really.

The longer answer: Kinda.

The bottom line

The eBook is hugely popular. I average around 400 visitors a day at http://ubuntupocketguide.com, peaking at 40-50,000 every now and again. The website advertises the print edition of the book, as does an advert within the PDF itself.

I’ve lost count how many people have downloaded the eBook but the last time I audited the figures, which was around six months after the book’s release, it’d seen around 500,000 downloads. I suspect that number has doubled since then. I encourage people to redistribute the PDF, including via BitTorrent, so auditing is practically impossible. (Redistribution is fine, but not modifying; the book uses a standard copyright.)

People like it when you give something away, especially in the Linux community, where practically everything is freely available. Big news, I know.

But how does all this translate to sales of the print edition?

Since going on sale at the start of 2009, the book has made me $9,000. Bearing in mind the book took three months to produce, that’s a salary of $3,000 per month, although costs such as hosting have to be deducted, and I also spent quite a few days marketing the book once published.

I’ve had worse salaries in my life, and I’m very grateful, but I know total royalties would probably have been higher had I gone through the traditional route of working with a mainstream publisher.

I estimate I have to give away 446 copies of the eBook for every sale of the print edition. Without Amazon S3, I simply couldn’t have done it.

The book’s been a success in terms of providing a free educational resource for the Ubuntu community, and I’m very proud of this achievement. But as a commercial endeavour, I wouldn’t recommend anybody take a self-publishing route.

If you want to earn money from writing, write magazine journalism. If you really want to write a book, get signed to a publisher. If you want an adventure then try self-publishing. Just don’t expect to do much more than cover your costs, though.

The future

The book is still chugging away each month, earning a little money. I’ll be a millionaire by the year 3000.

However, right now my big publishing experiment is $0.99 Kindle books. Take a look at my offerings. Maybe one day I’ll tell you how that went too.