eBook to Amazon Print-on-Demand: Eight weeks in

For the last four years, I’ve been selling various GMAT prep eBooks, such as my most popular resource, Total GMAT Math.  The advantages to the eBook format are enormous: Set-up costs were negligible, I never kept an inventory, customers received their orders almost immediately, and I kept a very high percentage of the sales price.

These reasons kept me from seriously investigating options that would allow me to produce print versions of my books.  Occasionally prospective customers would email, asking if there was a way to buy a hard copy, but often those inquiries would result in eBook sales anyway.  I don’t know how many sales I lost because of my lack of hard copy options, but I always assumed that the advantages of the eBook format outweighed the disadvantages.

Enter CreateSpace

A few months ago, I discovered CreateSpace, a subsidiary of Amazon, which offers print-on-demand services.  The most striking element of their service was the price: $0.  They offer a “Pro” plan for $39, renewing at $5 per year, which give syou better royalties.  If you can sell copies to the world outside of your immediate family, the Pro plan is a no-brainer.

Setting up a print version of Total GMAT Math through CreateSpace was laughably simple.  The eBook already existed as a PDF file (I originally wrote it in TeX), and CreateSpace needed a PDF file for the interior pages.  The designer who has made cover images in the past was able to easily generate a cover to the necessary specs.  The whole process took no more than a couple hours of my time.

Once you’ve entered some basic information (title, author, etc.) and the interior/cover files, they send you a proof copy.  (You pay the “member price” plus shipping.  I’m guessing that’s a little higher than their cost, but not much.  For me, it was $6.)  Mine was fine, so I ok’d the book and sent it out into the world.

Sales Channels

CreateSpace offers three sales channels: Amazon; an “eStore” hosted by CreateSpace; and an “extended distribution channel,” which allows libraries and bookstores to carry your book.  Royalties are different for each one.

Let’s get into the numbers a bit.  My print version of Total GMAT Math is a 7″ by 10″ paperback with 428 pages.  I sell the eBook for $49.95, so I set the retail price for the paperback at $59.95.  For sales through the CreateSpace eStore, I get about $42.  For sales through Amazon, I get about $30 (plus a couple bucks in affiliate income if the buyer comes from my website).  The extended channel is less; I’m not using that at this time.

(You can tinker with the specs and see the resulting royalties here.)

Given those parameters, I thought the eStore sounded like a great idea.  They allow some limited customization, so I got my logo on the page, but in the end, the eStore is pretty ugly.  I recognized that from the get-go, but for an extra $12 a book, I thought it might be worth a try.  At first, I sent prospective buyers from my website to the eStore.

It converted like crap. I think I sold one print book that first week.  My sales fluctuate, but this was in August (one of my biggest months), and I sold plenty of eBooks that week.  It’s not hard to see what held people back:

• I was sending people to a domain they weren’t familiar with.  The limited branding wasn’t enough to convey trust with my brand name, and who has heard of CreateSpace?
• CreateSpace uses an external checkout service (also unfamiliar), so by the time anyone places an order, they are on their third domain.
• Shipping is expensive and slow.  “Economy shipping” for this book is close to $5, and the estimated ship dates are something like three weeks into the future.  Faster shipping speeds still sound slow, and are much more expensive.
• Did I mention the eStore is ugly?

Long story short, it didn’t take long before I was sending my customers to Amazon instead of CreateSpace.  An extra $12 a book doesn’t matter if you’re multiplying it by zero.

Amazon.com

It’s been fascinating watching Amazon integrate my book into the site’s ecosystem.  At first, there was just a product page with the cover image.  Then the background information I provided.  Another few days, and there was an author page for me (also showing my Kindle book).  Around this point, users could “look inside” and view the full text of the book.

Within a few weeks, an early purchaser had posted a positive review and I had gussied up an author page with a bio, picture, and blog feed (Amazon’s Author Central makes this easy).  Then, once Amazon’s algorithms got a few customers to work with, I started seeing other books show up as recommendations, and just last week, the product page started offering “Frequently Bought Together” packages.

As soon as I started sending my site visitors to Amazon instead of CreateSpace, sales went up.  So far in October, I’m doing a little better than one print book per day — not even half of my overall volume for Total GMAT Math, but not an insubstantial amount, either.  Amazon rankings fluctuate a lot based on the last few hours of sales, but I seem to be hovering just around the top 100,000; when I get a couple of sales at the same time, I’ve seen my rank go as high as about #20,000, good for about #20th in the “GMAT” category.

(Be warned: As soon as you’re on Amazon and people are buying your book, that sales ranking is seriously addictive.)

Amazon Quirks

One frustrating bit is that Amazon persists in calling my book “Total GMAT Math (Volume 1).”   It’s not “volume” anything. It may have been my mistake when I initially entered the book’s metadata.  Regardless, I haven’t yet been able to fix it.  Amazon has a user-friendly way for authors to submit fixes like this–a few weeks go, I submitted a change request, and I got an email back a few days later, saying the change had been approved, and would take effect in 2-3 business days.  Now it’s been about 10 business days, and no change has been made.

The biggest surprise came the day that Amazon discounted the print book.  I set the print book retail price at $59.95 in part to encourage people to keep buying the eBook–after all, I get more money that way.  A few weeks ago, Amazon discounted the book to $46.81, where it’s been since.

This explains the big difference been the Amazon.com royalty and the CreateSpace royalty.  No matter what price Amazon sets, I still get $30.  In any event, it works out as a great deal for my USA customers–they get a price point cheaper than my eBook, and they get free shipping!

Some Early Lessons

I don’t do huge sales volume, so in this length of time, it’s tough to make concrete conclusions about the sales and revenue effects of making a print version available.  With that in mind, I can still draw some preliminary conclusions.

• Amazon converts. CreateSpace doesn’t. I don’t even really understand why CreateSpace, as an Amazon subsidiary, continues to offer their basic “eStore.”  The conversion rate more than makes up for the lower royalty rate.
• I’ve sold more units than I would have otherwise.  There isn’t a huge shift in revenue, but that’s up as well.  A lot of my business comes from selling additional products to happy customers, so more books sold–even for the same amount of total revenue–is almost certainly a good thing in the long run.
• Customers who have never seen my site are buying my book.  This is only a guess, but I’m surprised at how few of my book sales are generating affiliate commissions.  There are other explanations for that, but the Amazon ranking and recommendation systems do a good job of putting books in front of people that they aren’t explicitly looking for.
• Print-on-demand may not be internationally friendly. Again, I don’t have the figures to prove this, but it looks like almost all the eBook customers I’m “losing” to the print version are domestic.  That makes sense–it’s one thing to pay $0 shipping and wait a week; yet another to pay $10 and wait two weeks.

The early benefits have been enough to convince me to publish print versions of most of my other eBooks.  Clearly, a segment of my customer base prefers them, and at the very least, I’m not losing money by making them available.  There are probably also long-term benefits in having a presence as an author on Amazon.com, if only by making myself look more like an externally-validated expert.

Staten Island Half: Post-Race

1:47:37.  Not a half-marathon time I’m particularly proud of, but given how and how much I’ve trained for it, probably about what I deserved.

I can count a few small victories, though.  I achieved my primary goal of going out conservatively–at least relative to how I went out in my previous half.  My first mile was exactly 8:00, and the second was 7:38.  As it turned out, both were faster than my average pace, but I didn’t suffer because I went out too fast; my pace lagged because I wasn’t fit enough to race the distance today.

Also, I did a much better job of running tangents–according to my Garmin, I ran a total of 13.22 miles (versus about 13.35 in my previous half).  Given how much weaving I had to do in the first few miles, I’m not sure I could’ve done much better than that.  Finally, I pushed hard over the last mile or so, picking off plenty of folks and clocking one of my fastest mile splits of the race.

That’s about it for the positives.  I’m getting over a minor cold and suffered from some mild cramping through almost the entire race.  I was solid through the first five or six miles, but from there on, had to walk through some aid stations, and my running pace sagged.  As was the case in my May marathon, there’s some frustration here: My pace slowed down because I was tired, but I kept getting tired because I didn’t slow my running pace enough.  Once you run a few 7:30 or 7:45 miles, it’s tough to switch to a slower pace, especially when surrounded by runners moving at that 7:45 clip.

My lessons going forward are simple, and not terribly insightful.  If I want to get closer to a 90-minute half, I have to log more miles, and I have to do more long(ish) tempo runs.  A side benefit of more regular tempos, I hope, will be a better sense of pacing that will hold up even amidst a crowd.

Enough about me–here’s a bit about the race.

The Staten Island Half is a New York Road Runners event, and as such, it was extremely well-organized and well-staffed.  No complaints there.

I was one of a huge number of runners who started a bit late because of delays on the Staten Island ferry.  From what I overheard, it sounded like an earlier scheduled ferry hadn’t run, meaning that the 7:30 ferry was additionally delayed by the crowds.  And then well over one thousand runners arrived on Staten simultaneously with less than 15 minutes until race time.  No number of porta-potties is adequate for that kind of pressure.

The course was a bit disappointing.  It was almost entirely out-and-back, and there wasn’t much in the way of views, except at the start/finish.  I’m sure NYRR is forced to compromise–there are better spots on the island to stage 13.1 miles, but getting people from the ferry and back would be a nightmare.  Given the uninspired course and the transport to get there, it’s not a race I’ll likely do again.

Next up for me…

NYRR puts on a a 10k and a 15k in December.  My personal life between now and then is tilted toward things other than heavy, consistent training, so we’ll see where I’m at as we get closer.  I’m excited about the motivation to build some speed and see what kind of 10-15k times it translates into.

Two days to Staten Island Half-Marathon

Sunday morning, I’m running my first race since May, and my first half-marathon since March.  It’s only my second half overall, making it tough to set reasonable goals and decide on an early pace.

Back in March, I had been training regularly for about six weeks, in which time I logged roughly 235 miles pre-race.  Now, I’ve been running regularly for about eight weeks, in which time I’ve logged roughly … 230 miles.  Obviously not the same weekly volume.  Another major difference is that I’ve done next to zero speedwork this time around, whereas in February and March, I did two good interval sessions and two or three solid tempo runs.

The earlier training cycle got me to a 1:36:53, about a 7:25/mile pace.  That was better than I had expected going in; I think underestimated the benefit of running amidst a mass of humanity, much of which was going even faster.  If I had that race to do over again, on the same training, I have no doubt I could do better–my splits were all over the place and I made no effort to run tangents, so my Garmin said I covered about 13.35 miles–more like a 7:15/mile pace.

It’s clear that my recent training is not as good as it was seven months ago.  On the flip side, I have more race experience and an additional 800 miles in my legs.

Where does that leave me?  I’ve accepted that I don’t have a great shot at going under my 1:36:53.  If I ran it as a typical Sunday long run, I’d land at about 1:55.  There’s a lot of territory in between.

My foremost goal for Sunday is to pace myself conservatively.  Everyone knows they should go out slow, and then everyone goes out with the pack…way too fast.  I’m not going to do that.  I’m aiming for a first mile between 7:45 and 8:00, a second mile no faster than 7:30.  After that, I’ll have to see how my body holds up at that pace.  It leaves me plenty of time to clock 7:15 miles and come in around 1:37 … or settle in closer to an 8:00 pace and forget about beating my previous time.

Obviously, I have plenty to think about.  A reasonable secondary goal would be to stop thinking about it already.  🙂