[UPDATE: While the following deals with the problems created by CreateSpace’s Expanded Distribution, CreateSpace themselves have increasingly been using third-party printers to fulfill publisher orders. I have had numerous problems with my print orders, losing usually on average about 10% of my order to everything from crooked pages to machine roller smears. Further to CreateSpace, for their E.U. (and likely soon for their Canadian operations), it is not CreateSpace who are printing the books but Amazon themselves at their fulfillment warehouses. With so many potential printers of your product, the chances of poor or uneven quality have increased.]

Read any print-on-demand manufacturing contract and you will find a clause that states your POD manufacturer retains the right to subcontract the manufacture of your book to a third party, and as a consequence print quality cannot be guaranteed. This practice of subcontracting and its sometimes undesirable results are usually lost on the client who either a) never bothered to read the contract or b) doesn’t understand the implications. To put those implications into perspective, then, let me illustrate some key differences in the product produced by CreateSpace and that produced by Lightning Source.

I use both CreateSpace and Lightning Source International (LSI) for the print-on-demand manufacture and distribution of my novel, Baby Jane. When I received the proofs from each company, I immediately noticed two obvious differences: the CreateSpace cover had superior color fidelity and the font was truer in appearance. During the design of the book cover, I also had to adjust the spine width on the Lightning Source file as their system required a narrower spine even though both companies allegedly use 55lb crème paper. The result is that the LSI version of Baby Jane is 3mm thinner and 26 grams lighter than the CreateSpace version.

On the other hand, CreateSpace unfortunately use a print-to-laminate process for their covers, and excess humidity can cause the laminate to peel back from the cover stock. LSI still employ the superior process of printing to the cover stock then sealing it with a clear laminate.

All of which I noted in The Global Indie Author. However, when I received the proofs for The Global Indie Author itself, further differences in quality between CreateSpace and Lightning Source became blatantly apparent.

The Global Indie Author makes use of some thirty illustrations, mostly screen shots, which are difficult to translate into printable quality. But with some good advice (thank you Martin Evening) and a day spent in Photoshop, the images turned out surprisingly well. Still, I was apprehensive, so when the CreateSpace proof arrived and the illustrations looked great, I was thrilled.

And then came the Lightning Source proof. All I could think of as I looked through the mess of my book — which resembled a bad photocopy — was “What the hell?” The pictures had lost the bulk of their midtones, the lighter grays had washed out to white, and the darker grays had turned into a mash-up of black blobs. The image on the right below shows a scan of the CreateSpace proof while the left is the same page from the LSI version.

When the matter was taken up with LSI, they blamed me, stating that it was normal for anything below 15% gray to print as white and anything darker than 85% gray to print as black: I needed to adjust the tonal range of my images to conform to this alleged industry standard. But while I was used to losing some tonal range at the edges of the spectrum — 5% to 10% is not uncommon — to lose 30% of the gamut seemed excessive. And unfortunately for LSI, I happen to be a professional photographer with access to a worldwide network of professional photographers, and a quick post out to all and sundry resulted in a conclusion that 30% was indeed excessive, even by the lower standards of print-on-demand.

LSI then decided that I had merely received a bad proof, as tests at their end resulted in a better tonal range. Nevertheless, they warned me, with some 10,000 pages printed per hour — and all from various books — on a machine that is calibrated once per hour, they would not and could not guarantee the tonal ranges inside that 30% of the gamut; to be safe I would have to adjust all my images to conform to their specifications. That LSI had produced both a good and bad proof from the same file, and that there then remained the issue of me, the publisher, having to design to a moving target, was glossed over with LSI’s exhortations of their investment in million dollar “state of the art” equipment. To which I could only reply that they ought to find out what printing equipment CreateSpace is using because it is clearly superior.

I adjusted my images and ordered a new proof from LSI. The book, while much improved and acceptable for distribution, still did not equal CreateSpace’s: in particular, fine details did not print very well, rendering some of the text in the LSI images less legible than in their CreateSpace counterparts. At this point I took out the jeweller’s loop, and a closer look revealed the source of the differences. If you look at the image below, you can see how the CreateSpace printer (top) fills in the spaces between the black dots with shades of gray, while the LSI version (bottom) tries to create the same midtone using a black dot overlaying gray. The result is a blob of black with only a halo of gray, and a subsequent loss of fine detail. The next images illustrate this at even greater magnification.

This problem manifests itself in the text as well. The CreateSpace text is truer and sharper compared to the LSI text, which is thicker and softer due to the difference in printing techniques:

What made the situation with LSI all the more frustrating is that nowhere in their documentation is the client warned about the loss of 30% (30%!!) of the grayscale gamut; instead, one is simply told to produce an 8-bit, 300 ppi, grayscale image for interior files. When I queried them as to why they don’t warn their clients, I was told this was so as “not to confuse” us. Moreover, LSI argued, they receive very few complaints from clients, and even fewer from the client’s customers, the book buying public, so clearly it was just me.

But of course they don’t receive many complaints. The beauty for the print-on-demand manufacturer is that the client rarely if ever sees the end product. As long as the publisher–client is sent a good proof, they are blissfully ignorant of the variations in quality that may ultimately be sent to the consumer. The customer, in turn, has no original from which to compare, and will simply think your book ought to look like that, if regrettably so. And unless the customer writes a negative review or returns the book, the publisher happily remains in the dark, and it’s business as usual for the manufacturer.

So then, it has been asked of me, why not abandon Lightning Source and only contract with CreateSpace? The answer, unfortunately, reveals print-on-demand’s dirty little secret: the aforementioned subcontracting. For their Expanded Distribution, CreateSpace distributes through Ingram, and it is expedient to subcontract much of the printing to Lightning Source, which is owned by Ingram, to fulfill those orders. So the client–publisher receives that beautiful proof from CreateSpace and is blissfully ignorant that a customer outside of Amazon.com might be receiving a book printed by LSI. Worse still, as noted in The Global Indie Author and again in this post, LSI books are thinner, which means they need a cover that contains a narrower spine. But the file you sent to CreateSpace is sent to LSI, so the excess spine will have to be wrapped around to the front and back covers. If you use CreateSpace’s Expanded Distribution, then, you should design your cover to take into consideration the narrower spine that will result at LSI.

The problems outlined here are not confined to CreateSpace and Lightning Source — the practice of subcontracting happens across the board — and should be of concern to all publishers, not just indie authors: print-on-demand is the way of the future; it means less waste and less risk, which is of particular importance to smaller, independent publishers struggling to stay afloat. As the POD industry grows, it must address the hidden practice of subcontracting to third parties, especially where the quality is unequal or where the physical requirements of the book vary, and to be more honest with their clientele. It is misleading and unethical to suggest the client is receiving your product then rely on the fine print to cover the use of third-party printers, and hope we don’t become the wiser.

Update: I have been asked by authors for designing tips to avoid the worst of these issues. My suggestions are these:

1. Use white paper. I have noticed that the differences in book dimensions between LSI and CS are more pronounced when using crème paper.
2. Go to Lightning Source and on the top menu bar click on File Creation > Cover Template Generator. Fill in your book’s specifications and your email address. A template will be sent to you. Compare the spine dimensions of the LSI template and your CreateSpace cover template. That will tell you the likely difference between the two. Imagine the excess being pushed to the front or back of your book. Design your spine elements to take into consideration the loss of spine width (generally speaking, make spine elements smaller so they do not get cut off on the narrower spine), and your front and back covers to take into consideration some possible wraparound: since the excess will be cut off, increase the usual .25″ safe zone by whatever the excess amount is between the two templates. Note that the spine may be centered and the excess pushed to both front and back covers, or the cover may be properly centered and all the excess pushed back.
3. If your book has images, compress your tonal range to fit within the 15% to 85% gray.