by Alan Beatts

A few weeks ago I posted something about in-store print-on-demand machines that got a fair amount of attention and interest (Print-On-Demand Not Coming to a Store Near You).  The problem was that the only hard figures I could find for that post were based on two rather old articles.  In the interest of accuracy, I contacted On Demand Books, the manufacturer of the Espresso Book Machine, to get up-to-date information.  The result was a long and very pleasant phone chat with Jason Beatty, the company's Sales VP,  that clarified a number of things.

First off, I was wrong.  There does seem to be a financially viable way for large and mid-sized stores to have an Espresso Book Machine on site.  However, it is not based on the business model that I expected.

Before going into that however, I want to run down some of the basic costs associated with the Espresso Book Machine (henceforth "EBM") based on current pricing and actual installed usage.  Please feel free to skip these numbers if you like.

Initial Purchase
Machine Cost:  $101,000
Software License Fee:  $15,000 for 5 years (note 1)
Set Up and Training Cost:  $3,500 (note 2)
Total:  $119,500

Operating Costs (per book - note 3)
Content Fee:  $0.70
Transaction Fee:  $1.00
Materials:  $1.74
Maintenance & Utilities:  $1.83
Total:  $5.27

Average retail price per book is $14.55 (note 3).

Based on the math that I did in my first post I got the following figures -- At the rate of one book per hour, the profit in a year is $8688.  Based on the cost of the machine, I would pay off the machine and start making a profit in 11 years and 6 months.  At the rate of 3 books per hour, the machine pays off in 3 years and 10 months.  All those estimates are wrong since they were based on out-of date figures.

The new, more accurate, and up-to-date figures that I have work out this way -- At $14.55 retail minus $5.27 in materials fees leaves $9.28 profit per book.  Borderlands is open 8 hours per day, 362 days per year.  At the rate of one book per hour, the profit in a year is $26,847.88.  Based on the cost of the machine, I would pay off the machine and start making a profit in 4 years and 5 months.  At the rate of 3 books per hour, the machine pays off in a year and a half.

As you can see, based on the revised figures, it's a much better proposition.  But, there's another aspect to this whole thing.  What numbers of books do stores sell and what sort of books are they?

Books that can be printed on the EBM fall into four categories.  Academic titles, which are uncommon and quite expensive, public domain titles which are essentially free and provided by Google's huge database of titles, books from publishers, and "self-published" titles.

At this time, the number of titles from publishers is pretty limited.  The only one of the big six publishers whose books are available for the EBM is Harper Collins and, even in that case, the only titles are their back-list (i.e. titles older than about six months) trade paperbacks (not their hardcovers or mass market paperbacks).  All of the independent press titles printed by Lightning Source, which is the largest on-demand publisher, are also available.  But that leaves out the huge backlists of the other five big publishers (Macmillan, Penguin, Hachette, Bertelsmann (which includes Random House among others) and Simon and Schuster) as well as almost all of the new books published each season.  On-Demand Books is working hard to make deals with these publishers, but thus far without success.

So, what books actually sell and how many of them?

Based on the current installed base of machines, stores sell between 7000 and 14,000 books per year.  For Borderlands that would work out to between 2.4 and 5 books per hour, which is damn good.

But here's the surprise -- most of the books sold are neither public domain titles via Google nor are they in-print titles from publishers.  In the first year, 90% of the books printed by the current crop of in-store POD machines are self-published by customers of the bookstore.  In other words, someone comes into the store with an electronic file of their book, gives it to the store, and then the store prints it for them on the EBM.

The good news here is that there is an additional income stream from such self-published books.  Marketing materials from On-Demand Books suggest that most stores charge a set-up fee of around $150 per title to set up the file to print.  I'm not sure exactly where they got that figure but it seems to me that it would be hard to set such a fixed price.  Having a bit of experience with publishing and layout work (as well as lots of friends who run small presses), it seems that the likelihood of a "self-published" author coming in with a properly formatted file is slim (though that chance can be improved by creating a good formatting guide that authors can refer to before coming in with their file).  Consequently, I suspect that the stores offering this service charge a basic set-up charge and then an additional hourly design rate to get the customer's files into shape.  Stores also charge a per-book fee of $5 plus $0.045 per page to print each copy of the book.  That means the "retail price" of the book is right around $14.00 (assuming an average 200 page volume).  And that is on top of the fees for set up and layout.

The bad news is that 90% of the income from one of these machines comes from a process that is closer to running a copy-shop or a service bureau, not a bookstore.  It's not a process that most of the booksellers I know are well suited to -- moderately technical and involving potentially challenging customer service that is totally unlike bookselling (there's a world of difference between helping someone find the right book and getting someone's baby . . . I mean, their novel . . . to look right).

To be fair, the figures I've seen suggest that the sort of book printed changes over time.  Year one shows about 90% self-published but that number drops to 80% the second year, 70% the third, and 65% for the forth and fifth years.  But still, the actuality of having a machine in-store is quite different from the image of pressing a few keys and printing out a book from an existing catalog of public-domain and published titles.  Granted that does happen, but based on the figures, it happens only a couple times a day.  The rest of the time, it is a more complex process involving a non-trivial amount of time in front of a computer screen mixed in with time spent discussing details with customers and, perhaps most challenging of all, managing their expectations.  That's a fair amount of work.  What seems to me to be the hardest part would be orginizing the process, finding the right staff, and getting it all working smoothly.

But, if a store is willing to do that work, an Expresso Book Machine can clearly be a profitable investment -- especially if you have the right staff person who really enjoys the job and working with the customers.  However, without that work, it's even less attractive than my first post suggested.  If you cut out the income from self-published works, the figures are dreadful since, minus 90% of the sales, the pay-off time for the machine increases by a factor of ten (i.e. instead of the 1.4 year figure circulated by the manufacturer you're looking at something closer to 14 years!).  But that may change as more and more titles become available via print on demand.  If arrangements can be reached with more of the truly big publishers, even if only for their backlist titles, then the EBM and machines like it could live up to their potential.  Imagine being able to get a physical copy of any book in print at your cosy, local bookstore!  But, at least for now, the actuality does not seem to live up to the dream.

note 1:  This $15,000 is a single payment fee for a 5 year license for the software that makes the machine work.  At the end of 5 years (which is the expected life of the machine), you would need to pay again.  There is another option -- a yearly payment of $5000.  Obviously the math says that is a bad deal ($25,000 for a $15,000 license over 5 years works out to 13% interest annually).
note 2:  One cost not included above is the power requirement.  The machine needs two 220 volt circuits and one 110 volt.  This is not typical power service for a retail shop, which means that there will be a non-trivial cost to install electrical service.  For Borderlands it would run around $1500 to $2000, assuming our main electrical service would support it.
note 3:  Operating costs and retail price are on a per book basis in the first year.  The figures are averages based on actual usage of installed machines provided by On-Demand Books.