« Growth Markets, Resources

ISBNs and E-books: The Ongoing Dilemma


• Should publishers assign a unique ISBN for each e-book format of each title they publish, and what are the implications either way?

• There is no consensus of what is the best practice, but one thing is clear: you need to make the decision yourself and not leave it up to someone else.

By Erik Christopher

ST. CLOUD, MN: As more people venture into the e-book world, they inevitably come across a question they need to answer: Should I assign an ISBN to my e-book? But before we answer this first question, let’s look at why it’s important to designate a new ISBN for an e-book in the first place.

There are many reasons for using a unique identifier for an e-book. It helps with discoverability and allows you to separate out the different formats of the title, which in turn allows you as a publisher or writer to see how it is doing in various channels. The ability to measure the success rate of each e-book format for any given title is paramount to good marketing. The ISBN also gives you control over your title and your content. If you let someone else assign an identification number to your content, you lose that control, both in terms of quality and ownership.

Much of the debate about ISBNs and e-books stems from the use of the word “format.” The question being, when is an e-book in one format, say ePUB, no longer an ePUB? We know, for example, that a Kindle version is a new format, because it is no longer an ePUB file, but a unique version that will only work with Amazon’s software.

“A lot of it depends on where you will be submitting the e-book and who you will be working with for distribution,” says Laura Dawson, Content Chief and head of the Content Services group at Firebrand and co-chair of BISG’s Identification and Rights Committees.

Simply put, there are a variety of vendors and marketplaces selling e-books -– Apple, Amazon, Google (as soon as they come full force with it) Sony, B&N, Kobo and the list goes on. And some require ISBNs and some don’t care whether you have an ISBN attached or not.

Take Apple’s iBookstore, as just one example. If you submit your work to them directly -– they accept ePUB files –- they requires a 13-digit ISBN. They’ll add DRM to the e-book for you as well.

If you need to work in a different format, Apple suggests you go with one of their approved third part aggregators. These vary and they each may have their own stipulation for how to get the e-book to them. Smashwords works with the iBookstore to get your e-book there, but you have to start out with a Microsoft Word file and make sure it is formatted according to Smashwords’ specifications so they can convert into different formats.

“We currently only assign ISBNs to the ePUB format, since the retailers we distribute to who require ISBNs are taking our ePUBs,” says Bill Kendrick of Smashwords. “So, for example, if an author gives us an ISBN here at Smashwords, it will get attached to the ePUB that we generate here.”

Since Smashwords only accepts Word files, which limits the creative control a creator might have over his or her work, it essentially tosses the idea of an ISBN for every format out the window.

This doesn’t even go into all the other conversion houses and distributors out there that can get your e-book into all the aforementioned stores, and what they may require.

Next, consider this: what happens when you add a DRM layer on top of the original file, no matter the format? Does that signify a new creation of that work?

Again, let’s take an ePUB file for example. If you have the original ePUB file and you add DRM to it, is that another format and do you need a different ISBN for the DRM version and for the version without DRM?

Phil Madans, Director of Publishing Standards and Practices at Hachette Book Group, states: “In my opinion, which is also Hachette’s, no, it’s not a new format. The DRM is an anti-theft feature applied to the ePUB.”

Think of it as a bike lock on your e-book, it doesn’t change the bike really, just restricts what you can do with it. So, if you say that DRM is just a stop gap, then no new ISBN is needed. If you think it changes the e-book itself, then you may still want to consider one.

There are a variety of reasons to add the ISBN on your own as well, either as the author or publisher; one very important reason being that if you don’t assign an ISBN, the distributor could assign one for you, which could result in multiple ISBNs for the very same type of file (as sold through a wide variety of distribution channels). The bloat is likely to make collating and tracking sales data –- and thus looking at overall performance for a title — all the more complicated.

Though, as discussed here, there is no universal answer as to whether or not each format requires individual ISBNs, one thing is indeed clear: take control of the process. The worse thing you can do is lose control of your content or let another entity (whether conversion house, distributor or retailer) control the metadata and, accordingly, the invisible ties that bind you to your customers. Ceding too much control takes you out of the picture and makes this already complex situation all the more challenging.

Erik Christopher is owner of KC Educational Services, working as an independent publisher rep in the Midwest, and owner of Ugly Dog Digital, an all in one consulting, e-book conversion and sales company working with authors, publishers and libraries. Previously he was the National Sales Manager for North America at Blackwell, an academic library vendor for print and e-books. He is also Senior Advisor with Digital Publishing Partners. You can find him on LinkedIn, Twitter, and online at http://ebooknoir.wordpress.com/ and http://ebooknoir.tumblr.com/.

POLL: Do You Assign Individual ISBNs to Each E-book Format?

READ: Our earlier story “E-book ISBN Mess Needs Sorting Out,” say UK Publishers

This entry was posted in Growth Markets, Resources and tagged , , , , , , , , , , , , , , , , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Posted November 2, 2010 at 4:56 am | Permalink

    There are two initiatives to propose a way through this:

    The ISBN agency recommends that a publishers should assign a separate ISBN to each e-book format. They restate the principle that “a product needs a separate identifier if the supply chain needs to separately identify it”, and from this also encourages separate ISBNs for chapters if they are to be sold separately. See http://www.isbn-international.org/news/view/29

    To help associate different ISBNs with the same underlying content, the Book Industry Communiction (BIC)group have proposed the International Standard Text Code (ISTC). ISTC will be a new ISO identifier that identifies textual works and enables individual ISBNs comprising the same work to be grouped together for search and analysis. For more information about ISTC and how to register, go to the ISTC website. http://www.istc-international.org/html/

    I hope this is helpful.

  2. Posted November 2, 2010 at 7:29 am | Permalink

    Spinifex Press in Australia, a small publishing house, has been producing eBooks since 2006. We have applied new isbns to different formats e.g. kindle, ePub, PDF, DX Reader. It is useful since you can tell what format is selling. At this point in the technological changes, that is important information for publishers.

  3. Posted November 2, 2010 at 11:50 am | Permalink

    I have to disagree with Phil Madans’ comment that DRM is an anti-theft feature. It is much more than that as it dictates the user’s experience and usage features (e.g. printing, lending, text-to-speech etc.) An epub with different DRM may be the same format but is likely to be a different product from the customer’s point of view.

    Taking your bike lock analogy, even after you have unlocked it, it may be the same bike but will only go in one direction (Seattle or Mountain View) depending on whose lock you bought.

  4. Posted November 2, 2010 at 12:21 pm | Permalink

    Some additional context –

    When is a standard not standard practice?


    I am conducting the study for BISG (on this issue) and currently condensing my interviews into a findings report. I’ve conducted over 50 interviews with more than 70 people across the supply chain and thus should have an excellent perspective on what is current practice.

  5. Posted November 2, 2010 at 1:27 pm | Permalink

    I don’t believe in assigning an ISBN to anything. First of all, it is not a legal requirement in order to market and sell a book in any format, and second it is just a money making deal for Bowker. When an ISBN is assigned to any format of my books it is not by my choice but whoever I have to deal with to get my books onto the right markets. I deal with only one number for the printed book, and create or get my own files reformatted for ereading without the number being assigned at all. I use a unique catalog number to assign to each format of a specific book for inventory control purposes only. I rebelled against the ISBN system in the first place because it is completely arbitrary and does not allow my books into the bookshops anyway, because the bookshops ignore self-published books. So a number in this context is completely useless and just a money suck.

  6. Posted November 2, 2010 at 2:37 pm | Permalink

    As I told Mr. Cairns, your response to this issue is flavored by the size of your operation, and where you fall in the food chain. I would find assigning each format–Kindle, ePub, HTML, etc.–a separate ISBN onerous not just because of the cost but because it would further complicate an already complicated process of tracking sales across multiple vendors. Then there’s the matter of bundled formats. Do I assign yet another ISBN to the RTF file I submit to Fictionwise to be formatted into all of the major formats?

    One number per format seems to be a demand of the end-user distributors in the chain: libraries and bookstores. I can understand their desire to be able to see at a glance whether they have the format their customer/patron asks for, but I cannot believe there is no other alternative way to catalog than by ISBN, provided the ebook is, as we do here, assigned an ISBN for the “electronic” format.

  7. Posted November 3, 2010 at 12:33 am | Permalink

    I always thought that most publishers assign a single ISBN to the eBook. It makes record keeping easier for the sales are all linked to one number. This can get difficult when one has hundreds of titles.

    As for account sales, each of the vendors must report sales so it is easy for the publisher to see where they come from. That’s how a publisher gets paid.

  8. Posted November 5, 2010 at 11:43 pm | Permalink

    Similar to Michael Cairns, I have just completed an extensive study of multi-stakeholder behaviors, requirements, preferences and challenges relating to the issue of “ISBN per format” identification and meta-data issues pertaining to e-books.

    While I am not positioned to reveal the final report or study findings at the current juncture, what I can say is that this is neither solely a format, DRM or economics issue. Its much more, and in the aggregate ultimately boils down to what how any one stakeholder, a pairing of trading partners in the supply chain AND the CONSUMER who is buying that product need to differentiate a unique product and/or version of that unique product from another unique product and/or version of the unique product in cases where the unique product or version are made discoverable or available for purchase.

    For publishers, intermediaries, retailers, libraries and their respective supply chain trading partners, differentiating one product vs. another product and/or a product from its version assumes a number of business, functional and technical requirements and use cases.

    Quite differently from the physical book world where most stakeholders agreed and commonly implemented on the premise and benefit of standards like ISBN, in the world of e-books, these requirements (business, functional, technical) vary in so many dimensions from one stakeholder to another, and even in one particular stakeholder segment like e-book retailers, there is tremendous diversity of requirements.

    At the core, trading partners engaged in the cataloging and sale of e-books want to SELL e-books, and have taken the path of least resistance to sell an e-book. Some retailers don’t really “care” about whether or not the e-book format, which in many cases is in fact EPUB, is differentiated from another file format, let alone the same EPUB, because that’s all they are selling. In other cases, some retailers that sell EPUB also sell PDF versions of the same product or version of that product. Some retailers require an ISBN, but not on the basis of requiring a different ISBN for the EPUB vs. the PDF; some don’t require an ISBN at all, and instead have chosen different requirements for product identification — and some have even stated that if the publisher doesn’t assign or supply an identifier with their e-book, they’ll do it for them.

    In traditional (physical) supply chains, whereby cataloging, trading and sales reporting of one product and/or version of that product were facilitated by ISBNs assigned to hardcovers vs. paperbacks, or to the original edition and then separately to the 2nd edition, everyone agreed back in the 1970s that an ISBN was required for each different “binding” or “edition” — but a few major retailers made it a REQUIREMENT for publishers and trading partners to assign ISBNs to each different binding or edition, and if there was a conflict, there were problems in the supply chain. Point of sale systems got tripped up. People had to call each other. And once the ISBN got assigned properly, the book could then be sold. Sure, there were cases “after the fact” and that created problems in sales reporting but also when payments got rendered, but the ISBN played a role at the core to attempt to eliminate these issues and it worked because everyone seemed to agree and make it a requirement…but for the consumer, it was pretty easy to determine whether or not the book was different based on hardcover vs. paperback version.

    In the world of e-books, “bindings” were essentially equated to “formats” in the most current version of the ISBN Standard (2007). And when that version was authored and published, it was assumed th at the same file format was indeed the same product that would be catalogued, made discoverable and sold or lendable anywhere throughout the supply and demand chain. It was assumed that publishers and their trading partners would want to catalog and do sales reporting based on different file formats. And at that time, EPUB was…well…another file format. Did DRM exist then yes it did. Did additional secondary layers of DRM or integrations of DRM with things like username and password authentication to restrict file sharing being asserted by one or more retailers exist then? No (maybe in very limited cases). Did the iPAD or Nook exist then? No. Did the term “enhanced e-book” exist way back then in 2007? No. Did the Kindle for IPAD app or the Nook app for IPAD exist then? No. Did the Nook Color device vs. the straight-up Nook exist then? No. And with exception of Amazon Kindle, is every e-reader device, browser-based services like Google Editions pretty much supporting ePUB…yes. And today, there are multiple retailers using EPUB and the same DRM framework…But are there variances in the way the content is constructed in order to satisfy varying degrees of use on one device or another…well, in some cases…yes..and in some it doesn’t matter. For example, is the version of an e-book that was produced with HTML5 different than the one that uses Flash? Maybe…or in some cases, it would be important to know that when i’m buying the product….

    I think everyone’s in agremeent that a “file format” doesn’t necessarily determine whether a product and a variant version of that product are unique. Where the industry needs to quickly agree and continue to agree is on what makes a product or version of a product UNIQUE. What attributes does the end user need to know in order to find the e-book product or version of that product in order to satisfy his/her requirements or interests. And, any identification standard, whether its ISBN or another is going to serve common business, functional or technical requirements across trading partners, the standard as part of definition and within practice (Michael, I agree with you), must be aligned with these requirements in practice.

    The challenge here is that today, everyone seems to have a different set of requirements — particularly retailers — and contrary to some of the responses posted in this blog, some retailers don’t require ISBNs at all anymore for e-books…and the ones that do aren’t necessarily enforcing that the ISBN that is used for EPUB or another e-book file format is different (and can be the same in some cases)…so everyone rallying around a standard is going to be challenging in this new world, and that’s an unfortunate reality. And requirements change all the time, as does technology, not to mention new business models are emerging and raise deeper questions with respect to how and when to define a unique product (e.g., the time-delimited rental version of an EPUB vs. the EPUB version that has unlimited use.) Can identifiers solve that problem alone? No, and most certainly not on the basis of format vs. format, let alone when meta-data contructs and requirements vary —

    But at the same time, if I’ m a publisher who at minimimum is selling an ePUB format AND a PDF format of the same book across multiple retailers, I’d like to easily track, distinguish and compare sales of the product by version, and in some cases I’d want use a format delimitor to help me do that (EPUB vs. PDF)…I’d like the end customer (the consumer) to know what they are buying and how they can use it (or not) on multiple devices, particularly in cases where I may have created or had someone else create a variant version of that EPUB file format (2 EPUB versions…)…and when the price of an e-book version is changed dynamically with one retailer vs. another, or with 3 retailers but not the other 4 that are selling, some might even argue that the 9.99 version of that product that was once 12.99 days before is…well…a different product—is that the job of the identifier, the meta-data…both to solve?

    In addition to requirements, these issues also need to be addressed from use case perspectives…for example, after the e-book has been published for IPAD and Nook and the publisher announces a new way to read that e-book on their new facebook app and share comments with friends, one might argue that this is…well…a different product from the previously released version which didn’t have the Facebook compatibility integration code built into it…okay, this may sound out of bounds for some people reading this but, again, what makes a product unique from another???

    Lets boil this down in more “practical digital” terms…As a consumer, when I buy a movie on iTunes, I know in most cases, when I’m buying the product, its clear that the “high-def” version of the movie is different from the “low-def” movie. I know that if I bought the “low-def” version, it won’t appear in high-def on my high-def TV or any high-def TV for that matter…like if I bought the low-def version to my friend’s house to watch it on his high-def TV, everyone on the couch would be like “dude, WTF…why didn’t you get the high-def version”…and start calling me Heidi…(football fans u might get that joke)…and the movie studios making these products and versions are assigning different STANDARD identifiers (ISAN) which has made life a lot easier for those selling their products…

    Or in the case of a Blue-Ray movie, if I’m buying blue-ray version, I know while I’ll be able to watch it on my Playstation 3, that I WON’T be able to watch in on my standard DVD player in my bedroom. I can buy that blue-ray from many retailers in most cases…

    I also know that if I If i’m a retailer who is making a publisher’s product available on 2 different e-reading devices, particularly if there is a material difference between one version vs. another (Nook vs. Nook Color), unless the file has been constructed to adapt irrespective of the device its sold on, i’d like to think that the retailer and the publisher would think of the consumer who has the color device, expects the e-book to be “in color” and then gets a black and white version…with the same identifier, or without meta-data that distinguishes the two, its the consumer who is compromised in these cases — and that BOTH the publisher and the retailer would want to track these products individually — just some of the cases that will only get more complex and dynamic over time as this all evolves and plays out.

    In my view, and at least until the technology and business model issues shake out, or until everyone can agree on how a standard should ultimately be structured to meet everyone’s needs, the rule of thumb should be that if the e-book product or version needs to be distinguished in the mind of the consumer on the basis that it will impact how the consumer can OR expects to use it, it should be identified uniquely and described transparently no matter where, how or when that consumer acquires it — but not on the basis of just file format or DRM.

    Whether the ISBN or any other industry standard for identification for that matter can ensure this happens is indeed an important question that the industry may not so easily answer, but deserves clarity and guidance.

    And while some publishers might feel like ISBNs are too expensive to buy and assign to multiple versions of an e-book product (especially EPUB), and that having to manage multiple meta-data records for each version of a product for which they believe are only one in the same product despite carrying defining, differentiating attributes…NEWSFLASH…the end consumer deserves the right to know before they click the buy button and reach the point of “no return”…so for those publishers that are indeed creating multiple EPUBs or PDFs, or enhanced versions of an e-book that don’t “work the same” on multiple devices, so as long as there is a clear business case for doing so, you might very well find it easier, more practical and beneficial in the long-run to assign a separate ISBN to each version. And to be clear, this is ALL coming from the guy who no longer runs the US ISBN Agency at Bowker but does continue to have a vested interest in helping the publishing ecosystem and the consumers it serves to get it right.

    Comments, debates, rejections, antecdotes, solutions, ideas are all welcome!

  9. Timothy
    Posted November 12, 2010 at 10:06 am | Permalink

    Thanks, Andy, for the great post. Covers a lot of great issues. I enjoyed reading it. Enjoyed meeting you at TOC this year.

  10. Posted November 12, 2010 at 10:54 am | Permalink

    Andy- Thank you for your thoughtful post. Helped me formulate PDF v. ePub ISBN strategy.

  • Get Publishing Perspectives in your inbox each day and stay up-to-date on international publishing.