INDUSTRY STANDARDS

Part 1:

Why, How, and What Booksellers Need You to Tell Them

Especially if you’re new to publishing, you may be familiar with ISBN but bewildered by industry terms like BISAC, ONIX, EAN, imprints, and edition type. And no matter how long you’ve been in the business, when you’re struggling to get a book edited, designed, and off to the printer, it can be hard to remember that your wholesaler, distributor, and major retailers need all the information these terms cover at least six months before a book goes on sale.

These aren’t mere details. They are what determines whether a book moves from warehouse to retailer to customer without confusion or delay. It’s also important that you as the publisher keep this information updated when a project is postponed or abandoned, when a title or a price changes, and when a new edition comes out.

Without accurate book data, you’re likely to lose sales. A bookseller may tell a customer that a title is out of print, not knowing that it can be special-ordered. Or a cataloger serving the visually handicapped market may be unaware that you’re now offering a large-print edition of a popular title. And without accurate on-sale dates, wholesalers and booksellers may not place orders in time to have books available when publicity starts for a title.

In this article, and in articles coming up, we’ll define book data and how it relates to ONIX, explain why wholesalers, distributors, and retailers need all this data from you; what the ill effects are when information is inaccurate or incomplete; and the options for submitting data, either through vendors or by doing it yourself.

Publishers issuing fewer than 100 titles a year can submit their book information via bowkerlink.com to Bowker, which publishes the industry standard for bibliographic information, Books in Print. Larger publishers and those who are tech-savvy can submit information via ONIX or with Bowker’s Excel template. For details, see bowker.com/index.php/data-file-submission. There is no charge to submit to Bowker.

ONIX is not data, but a format for sending data about your books. The international standard for representing and communicating book industry product information in electronic form, it was developed and is maintained by EDItEUR together with standards organizations such as the British group Book Industry Communication and the American Book Industry Study Group.

You can find detailed directions about coding for ONIX on the Book Industry Study Group Web site (bisg.org) in the PDF titled “Product Metadata Best Practices for Data Senders.” What follows explains why the information is necessary for book metadata.

The Data You Need to Provide

ISBN. All new titles should be assigned a 13-digit ISBN, and that should be submitted in data feeds, without spaces or hyphens, 180 days prior to a title’s on-sale date. As BISG points out, “The book industry supply chain is almost completely dependent on the ISBN numbering system. Transmitting an accurate product identifier for every item is the only way a publisher can ensure that its trading partners will order the correct products.” You must submit a unique product identifier for every single product.

Title. This is the complete name of a published product, including the subtitle if there is one, as it appears on the title page.

You may use a variation of the title on dust jackets or spines, but don’t submit variations in data records.

Titles should be presented in the appropriate title case for the language of the title, which is defined for English-language books as headline style. In other words, “the first and last words and all nouns, pronouns, adjectives, verbs, adverbs, and subordinating conjunctions (if, because, as, that, etc.) are capitalized. Articles (a, an, the), coordinating conjunctions (and, but, or, for, nor), and prepositions, regardless of length, are lowercased unless they are the first or last word of the title or subtitle.” Titles published in Spanish and French should have the first word of the title and of the subtitle, and all proper nouns, capitalized. All other words should be lowercase. Titles should never be presented in all capital letters as a default.

A title, even if it is only a preliminary title, should be supplied 180 days prior to the on-sale date. Preliminary or working titles should be updated to final titles at least 120 days prior to the on-sale date.

Contributor(s). These are the names, titles, and roles of everyone or every organization named as a participant in the project.

Each contributor, along with his or her role, should be listed separately. Contributors may include authors, illustrators or editors, or the editors of a publishing company.

Publisher/Imprint/Brand Name. A publisher is defined as the entity that owns the legal right to make the book or other product available in this form. Publishers may be incorporated businesses, divisions of larger companies, governmental agencies, nongovernmental organizations, educational institutions, or individual persons.

Corporate names should omit any suffixes denoting incorporation (e.g., Inc., Ltd., S.A.). Names should be presented as they normally appear in print (e.g., Alfred A. Knopf). The imprint is the “brand” name that the publisher uses on the title page of the book. Imprint names usually also appear on book spines and dust jackets. For example, Vintage Books is an imprint of Alfred A. Knopf. BISG guidelines specify that imprint names should not indicate their parent publishing companies (e.g., Checkmark Books, An Imprint of Facts on File); publisher names are listed separately.

When you submit publisher and imprint names, do not include copyright, trademark, or other symbols, because these can cause problems in searching and indexing names in bibliographic database catalogs.

Like the data elements described earlier, this one should be submitted 180 days—six months—prior to the on-sale date.

Price. This means the suggested retail price, expressed in dollars for titles to be sold in the United States. It is required even if the price is not printed on the book.

Publisher’s Proprietary Discount Code. Even if you don’t offer different discount schedules, you should complete this field in your data submission. BISG specifies that this code cannot be the value of a percentage discount from the retail price. It can, however, be a code that refers a customer to a list of discounts that you have on your Web site or in your trade catalog.

Publisher Status Code. This code tells buyers where the book is in its product life cycle, and it should be submitted for every book you have published or plan to publish. For new books, it should be submitted 180 days before the on-sale date.

The available codes to use are:

• Cancelled: The product was announced and subsequently abandoned.
• Forthcoming: Not yet published; your expected publication date should be included.
• Postponed Indefinitely: The product was announced and subsequently postponed with no expected publication date.
• Active: The product was published, and the publisher is accepting orders for it, though it may not be immediately available.
• No Longer Our Product: You’ve transferred ownership of the product to another publisher.
• Out of Stock Indefinitely: The publisher is no longer accepting orders for this title, although stock may still be available elsewhere (on bookstore shelves, for example), and there are no current plans to bring it back into stock. This code does not specify whether returns are still being accepted.
• Out of Print: The publisher will not accept orders for this title, though it may still be available elsewhere as a new book, and it will not be made available again with the same ISBN. Using this code normally implies that the publisher will not accept returns beyond a specified date.
• Inactive: The book is now permanently or indefinitely unavailable in the sense
• that the publisher will not accept orders for it, though there still may be some stock available elsewhere. This code can be substituted for “Out of Stock Indefinitely” and “Out of Print.”
• Unknown: The use of this code is discouraged; trading partners expect publishers to know the status of each of their titles.
• Remaindered: The book is no longer available from the current publisher, under the current ISBN, at the current price. It may be available through another channel such as a remainder dealer’s catalog.

Product Availability Code. You should specify this using one of the following codes:

• In Stock: Available from the publisher as a stock item.
• Awaiting Stock (i.e., on order): Not yet available, but will be a stock item. BISG recommends this code for books you are importing, when they have been published in the country of origin but have not yet arrived in your country.
• In Stock: Available from the publisher as a stock item.
• To Order: Available from the publisher by special order.
• Manufactured on Demand: Available from the publisher by manufacture on demand (today, usually POD).
• Not Available: Not available from the publisher.
• Not Sold Separately: Must be bought as part of a set.

Product Form (Format/Binding/Packaging). Although most of us talk in terms of the paperback edition or the large-print edition of a book, BISG clarifies that these are really book formats. Examples include:

• trade paperback book
• mass-market paperback book
• hardcover book
• audiobook on cassette

Publication Date. There is no single definition of “Publication Date” in the U.S. book trade, so publishers can choose whatever pub dates they like. BISG Best Practices advises: “Publication Date is defined by many key accounts in our market as the date on which a retail consumer may purchase and take possession of a given product,” typically the date on which a book is on sale in bricks-and-mortar bookshops. “Where a book is sold via online or mail order prior to its appearance in physical stores, the publication date is defined by many key accounts as the date the consumer will receive the book.”

On Sale Date (or Strict On Sale Date). Books like the Harry Potter novels are often embargoed for sale until a certain date. This is the On Sale date.

BISAC Subject. The BISAC subject headings, which describe the topic of a book and are often printed on the upper left corner of the back cover of the physical book, and transmitted electronically to online stores, help bricks-and-mortar stores shelve titles and online retailers categorize their data.

The current list of BISAC Subject Headings consists of approximately 3,600 subjects grouped in 50 major categories. For more information, see bisg.org/publications.html.

Language of Product Content. Every applicable language that is used for a significant portion of the book content should be indicated.

Series. This can be any number of books that are published over any time period and grouped together, usually for marketing purposes. Some publishers offer standing orders for each new publication in a series. A series does not usually have its own ISBN, EAN, or UPC, and it is not usually sold as a single item (as a set might be). Some books belong to more than one series.

Series Number. If books in a series are sold in successive order, this is the number of an individual publication. Many books in series do not have series numbers.

Edition Number. An edition number is required for a numbered update of previous publication. First editions do not need edition numbers.

Edition Type/Description. When you publish a version of a work that is “materially different” from an earlier or simultaneous version, you are creating a different edition. As BISG explains, however, “The same work is often published simultaneously in hardcover, audio CD and audiocassette, and that work will subsequently be published in trade paperback and possibly mass-market paperback as well. These variations in product form do not constitute different editions.”

Examples of edition types are:

• Abridged: Content has been shortened.
• Adapted: Content has been adapted to serve a different purpose or audience, or from one medium to another (for dramatization, for example).
• Annotated: Content is augmented with notes.
• Braille: Published in Braille.
• Critical: Content includes critical commentary.
• Coursepack: Content was compiled for a specified educational course.
• Enlarged: Content has been enlarged or expanded from that of a previous edition.
• Expurgated: “Offensive” content has been removed.
• Facsimile: Exact reproduction of the content and format of a previous edition.
• Illustrated: Includes extensive illustrations that are not part of other editions.
• Large Type/Large Print: Printed in 14-point or larger type.
• Microprint: A printed edition in type too small to be read without a magnifying glass.
• Media Tie-in: Published to coincide with the release of a film, TV program, or electronic game based on the same work.
• New Edition: Where no other information is given, or no other coded type is applicable.
• Revised: Content has been revised from that of a previous edition.
• School Edition: An edition intended specifically for use in schools.
• Special Edition: Anniversary, collectors’, deluxe, gift, limited, numbered, or autographed edition.
• Student Edition: When a text is available in both student and teacher’s editions.
• Teacher’s Edition: When a text is available in both student and teacher’s editions; also used when instructor’s or leader’s editions have different material.
• Unabridged: When a title has also been published in an abridged edition; also for audiobooks, regardless of whether an abridged audio version also exists.
• Unexpurgated: Content previously considered “offensive” has been restored.
• Variorum: Content includes notes by various commentators, and/or includes and compares several variant texts of the same work.

Volume Number. This indicates the number of a particular publication within a set and the total number of publications in the set.

ONIX Audience Code. Because you should supply only one audience code for a book, use the code for the primary audience even if you believe the book will appeal to several audiences.

Audience codes include:

• General/Trade: For a nonspecialist adult audience.
• Children/Juvenile: For a juvenile audience, not specifically for any educational purpose.
• Young Adult: For a teenaged audience, not specifically for any educational purpose.
• Primary & Secondary/Elementary & High School: Kindergarten, preschool, primary/elementary or secondary/high school education.
• College/Higher Education: For universities and colleges of further and higher education.
• Professional and Scholarly: For an expert adult audience, including academic researchers.
• ELT/ESL: Intended for use in teaching English as a second language
• Adult Education: For academic, vocational, or recreational courses for adults.
• Age Range of Target Audience: It’s recommended that you specify the exact age in years or school grades when the intended audience comprises children or young adults. Age ranges such as “up to age 5” and “age 8 and older” are discouraged.

Case Pack/Carton Quantity. The number of units of the book or other product that are packed in the product’s standard shipping container.

Replaces/Replaced By (or Related Product). Here is where you indicate the product identifier (usually an ISBN) for a previous edition of a current product and the identifier for the successor edition of the same or similar product

Territorial Rights. This tells retailers and other resellers where they can sell your book, and it protects the rights of those to whom you have licensed or sold specific territorial rights. For example, you might license rights to a French edition to a Canadian company for sale only in Canada.

Bar Code Indicator. Specify what kind of bar code the book carries, usually EAN or UPC, and where the bar code is positioned.

Weight and Dimensions. Used only for physical books, this describes the length or height as the measurement of the spine from top to bottom; the width as the measurement perpendicular to the spine; and the depth or thickness as the measurement across the spine of the book from left to right.

Return Code. Here is where you describe your general returns policy. Special returns conditions (for example, for customers who have received deeper discounts because they were buying on a nonreturnable basis) should be described elsewhere.

Examples of returns codes are:

• Yes: returnable, full copies
• No: not returnable
• Conditional: contact publisher for requirements
• Strippable: can return stripped copies instead of full copies

Page Count, Running Time, and Extent. Unless your book has no numbered pages, this is not the total number of pages in the book. Instead, it’s the total sum of the numbered pages. Books that have pages numbered in both roman and Arabic numerals should count the total of both. For multivolume books sold under a single product identifier, use the page count for all the volumes combined. If the individual volumes are sold separately, each product record should carry a page count for that volume.

Running time means the total length, in minutes or hours, of recorded content.

Distributor/Vendor-of-Record. This is the company that takes and ships customer orders. As BISG explains, most publishers designate one vendor-of-record for each geographic rights region or market segment.

Some vendors-of-record will service multiple geographic rights regions and/or market segments. For example, you might contract with one wholesaler to fulfill orders to general trade bookstores in the United States, and a different one to provide this service in Canada. You might also designate a third firm to fulfill orders from Christian bookstores and yet a fourth to fulfill orders from newsstands and other mass merchants.

According to BISG, wholesalers, such as Ingram, Baker & Taylor, and Brodart, “should not be described as a vendor-of-record if they are simply reselling a publisher’s products. Only if a wholesaler is a publisher’s designated vendor-of-record should a wholesaler be listed as the vendor-of-record in an ONIX message. “

Number of Pieces. The number of saleable components composing a single product. For example, an audiobook may consist of 10 CDs, or a gift product may consist of a book and a toy.

This term applies to prepacks, dump bins, and counter displays where a single product includes several saleable pieces.

Textual Description of Product. This descriptive text is similar to (or perhaps the same as) text printed on the flap of a dust jacket or on the back cover of a book or DVD package.

Illustration Details. These list how many and what kind of illustrations or other images are in this product.

Acceptable codes to use are:

• Unspecified
• Illustrations, black & white
• Illustrations, color
• Halftones, black & white
• Halftones, color
• Line drawings, black & white
• Line drawings, color
• Tables, black & white
• Tables, color
• Illustrations, unspecified
• Halftones, unspecified
• Tables, unspecified
• Line drawings, unspecified
• Halftones, duotone
• Maps
• Frontispiece
• Diagrams
• Figures
• Charts
• Printed music items: printed music extracts or examples, or complete music score(s), accompanying textual or other content

Digital Image of Product. This is a digital photograph or scan of the product suitable for display on Web sites. Today this data element is mandatory. It should be named by an ISBN, EAN, or item-specific UPC and submitted as a TIFF or JPEG, scanned at 150 dpi and in RGB.

The longest side of the digital image should be at least 750 pixels, with the shorter side proportional. Book images should be a flat front cover scan cropped tight to the sides of the product. In cases where the front cover image is of little merchandising value, publishers should also supply a back cover image and/or an image of the title page of the book.

Warning

If you don’t submit the right data in the right form at the right time, you’re likely to lose sales at several points and for several reasons.

Those of you who aren’t yet motivated to start—or keep—providing good data are advised to hold on to the instructions in this article; the next installment in our data-quality series will provide more motivation by focusing on the ill effects of ignoring them.

And for those of you who find all this too daunting, we’ll also be providing information on hiring help.

Beyond Bowker

To ensure that your titles are listed correctly and completely with wholesalers and major online and offline retailers, you may also want to submit this same information to them directly.

Some will want additional information, such as author bios, photos, and, for juvenile books, the Lexile score. For specifics, check these Web pages:

Amazon.com: “Books Content Update Form,” reached from “Publisher & Vendor Guides,” amazon.com/gp/content-form/?ie=UTF8&product=books.

Baker & Taylor: “Vendor Title Submission Form,”

edi.btol.com/PublisherSubmissions/Pubinfo.aspx.

Barnes & Noble: “How to Submit Content,” reached from the “Help Desk,” barnesandnoble.com/help/cds2.asp?PID=8150&.

Borders: “Customer Care,” borders.com/online/store/CustomerServiceView_borderscommunity; recommends that publishers contact Bowker and wholesalers.

Ingram Book Co.: Email bookbuyer@Ingram.com for details if you do not have an account and buyer established.

Online Computer Library Center: “Information and Services for Publishers,” publishers.oclc.org/en. If you’d like more information regarding subject headings, this page provides a link to a BISG Webcast, “BISAC Subject Headings: Connecting Books and Readers.”

Part 2:

Don’t Let Bad Data Cripple Your Sales

If you’re like most publishers, you check on how retailers are marketing your books online, and on how many libraries have purchased copies. Suppose you type a title’s ISBN into the Google Images search box, and up comes the cover of someone else’s book. Or Amazon.com shows the wrong author for your newest title. Or someone phones to say the local bookseller can’t find that new title in the Ingram database.

Chances are, these problems were caused by erroneous or incomplete information—or “metadata”—about the book, maybe a typo in the ISBN you used as a JPG’s name, for example, or a glitch in submitting the author’s name to Bowkerlink.com that kept the book title out of a database.

“Desperately Seeking Good Data: Why, How, and What Booksellers Need You to Tell Them” provided definitions of the metadata elements you need to provide. This month, the focus is on why it’s so important to submit correct information on time to the companies that maintain bibliographic files, and on how you can make changes or correct inaccurate data (see “Metadata Nitty-Gritty,” below).

There’s a very simple reason distributors, wholesalers, and retailers need two or three dozen pieces of information on every one of your titles: that information is what ensures that the correct edition of your book gets through the supply chain from your warehouse to the customer in the desired format. Given the volume of books published today, and the fact that the title of a book cannot be copyrighted, a customer can’t order by title alone. When I asked my local bookseller for Jane Austen’s Emma, her database brought up 144 entries. Similarities among a publisher’s ISBNs mean that confusion over a digit or two could result in the wrong book being shipped—and then returned.

The additional information that is often requested has another purpose: to help a book sell when pictured on Web sites and in catalogs. That’s why publishers are asked to provide cover and back cover images, information on images, contents pages, and author photos and bios. Advertising schedules, author tours, and the like, often listed as “optional” on content submission forms, help booksellers of all kinds determine the size of their initial orders, how to display a book, and what special promotions to do, whether online or in a storefront.

What Squelches Sales

A few of you might be mentally protesting that none of this matters because you’re only doing books through digital printing/print-on-demand programs so that you can sell them in your storefront, or to your association members, or when you make presentations. If you’re planning to produce books through such vendors as Amazon.com’s CreateSpace or Ingram’s LightningSource, though, much of the metadata is still required—and if you don’t submit it as specified on your vendor’s Web site, your project will be delayed.

A couple of years ago, when he was speaking to the Evangelical Christian Publishers Association, Barnes & Noble’s Joe Gonnella pointed out what happens when metadata is incorrect or incomplete. Gonnella (B&N’s vice president of inventory management and vendor relations) and his colleague Marcella Smith (the B&N small press and vendor relations director) tell me these problems still exist.

If anything, they’re getting worse. As Gonnella explains: “At any given time, Barnes & Noble has $4 million in purchase orders outstanding on NYP [not yet published] titles more than 30 days past publication date. These orders will cancel after 90 more days. When we cancel these POs, we have to cancel any customer orders we might have taken via our Barnes & Noble.com presale program.”

Translation: That’s $4 million in sales that could be lost—at Barnes & Noble alone.

But that’s not all. Millions of dollars in backlist sales are also lost, Gonnella continues. This happens when:

U.K. or other international titles that were once imported are out of stock and not replenished.

Booksellers don’t know where to send purchase orders because distribution arrangements have changed and ordering information has not been updated.

There is no information about a book’s status in the metadata, and, without information about whether the book is still in print, no possibility of satisfying a customer’s request.

Gonnella offers three examples of how inaccurate or incomplete book data can cause delay, expense, and lost sales.

First, if publishers don’t provide a book description and a table of contents, a search for a subject that is not explicitly mentioned in the title won’t bring up the book, “squelching any possible sale to that potential customer.”

Second, if delivery of a backlist title slips more than 30 days past the on-sale date you’ve entered into metadata, Barnes & Noble asks customers if they want to cancel the order. Typically, they do.

Finally, Gonnella and Smith emphasize the importance of a piece of information that many publishers overlook: the quantity of books per case. Especially when a title is being rushed to market before a media event, the initial orders could be done by case, saving countless hours of labor in Barnes & Noble warehouses.

“Publishers should know the carton quantity at least six weeks in advance of the on-sale date,” Smith points out, encouraging publishers to evaluate carton pack in terms of easy fulfillment of reorders.

The quantity of books in a carton is also important to the regional wholesalers I consulted. It’s easier to order, and to warehouse, books in full cartons. And carton pack is an easy number to obtain from your book manufacturer. Once a book is in production, the spine width will have been calculated, the printer can tell you what its usual carton size is, and you can specify a different size if you need that and your budget allows. When I published and distributed job-search guides, making lots of deliveries myself, I always asked that cases not weigh more than 30 pounds, and I based my discounts on case pack, to encourage orders of that quantity.

Duse McLain, whose Thistle Press in Bellevue, WA, sells pocket-sized guidebooks for walking tours, often to tourist-attraction gift shops, told her printers to reduce the number of books per case because few customers were willing to buy 140 books at a time.

Providing inaccurate data means running still more risks. Especially with a backlist title, it might mean the book would be dropped from the store’s inventory. “If a book appears to be unavailable for several months due to us not having updates on its status, the buyer may replace it with another title,” Smith notes.

Even if a bookseller keeps the title in its database, difficulty in obtaining stock will affect the book’s sales—and that affects how its inventory level is modeled. At Barnes & Noble, models are based on anticipated sales of a title in a category in a store, to ensure that Barnes & Noble has what it considers the ideal inventory level for a particular branch.

How Good Data Helps You

On the upside, providing complete, accurate book data on schedule increases the chance of your books being better promoted by booksellers. As Smith explains, chain stores develop their merchandising and promotion schedules at least six months in advance. Barnes & Noble buyers need that time to review books, get them in the database, and place orders. Even though most of us don’t have the good fortune to have our books selected for chains’ advertisements, getting our information in 180 days in advance of publication gives stores like Barnes & Noble a chance to have books in stock when our publicity begins.

If you remember the 30-some pieces of information last month’s article specified as needed even in a basic Bowker entry, you may be feeling overwhelmed or exasperated by this point. All this information doesn’t come your way all at once in a tidy package. That’s true, Smith acknowledges, and she emphasizes that she doesn’t need the information all at once. “Send us your data as you receive it,” she says. “Don’t hold it to send it all together.”

Wendall Lotz, vice president of metadata, Ingram Book Co., underlines the value of timely data in making sales. Getting text 180 days in advance and images 120 days in advance at Ingram is especially important in library sales, he says. “We need to place orders with publishers four or five months in advance of the on-sale date to be able to respond promptly to library purchase orders,” he told me, because when library acquisitions staff see books reviewed, they want to place orders immediately.

Theoretically, Ingram could process a book in as little as a month (receiving the metadata from the publisher, forwarding it to buyers, getting the buyer to place an order, having the order received by the publisher, the books shipped to Ingram, and then received and processed at Ingram warehouses), Lotz says, but if a title had to be processed so close to its on-sale date, it could not be included in any of Ingram’s many promotional programs, many of which do not involve extra fees to the publisher.

“If you can’t find it, you can’t sell it” is one of Lotz’s maxims, and he described how Ingram tries to find titles that its customers want. It tracks what’s on Oprah, what’s on PBS, what’s on NPR, and what movies have tie-ins, so that booksellers can get those titles. (As he and so many booksellers have noted, a customer may ask for a book seen on Oprah but remember only the author’s name and the color of the cover.)

Ingram’s most important new tool for bookstores is the Ingram Wire, Lotz says. Launched in December 2009, it includes a downloadable desktop application with stock information. Booksellers can also use the app to receive alerts on top awards and breaking events. The titles in the app link directly to Ingram’s iPage. Publishers whose titles are receiving such national publicity can arrange to notify Ingram as often as daily of upcoming media appearances and tie-ins to keep metadata updated.

Metadata Nitty-Gritty

Convinced that metadata is vital to the sales of your titles?

Once you have the 30-some pieces of information outlined as necessary by the Book Industry Study Group (see “Desperately Seeking Good Data: Why, How, and What Booksellers Need You to Tell Them” in the April 2010 issue, archived at ibpa-online.org) and any additional information that distributors, wholesalers, chain-store buyers, and the trade press want, you can start by checking what’s online for your existing titles and for any forthcoming titles you’ve submitted to Bowker.

If you don’t already have your book titles entered for Google Alerts (google.com/alerts), spend a few minutes with this and other search engines. They will quickly show you what data is being used by online retailers, libraries, and reviewers.

Then, if you spot inaccuracies or missing information, the logical next step is contacting Bowker via bowker.com/index.php/data-file-submission. If Bowker has correct data but you’ve found errors elsewhere, check the April issue of the Independent for appropriate contacts at Amazon.com, Baker & Taylor, Barnes & Noble, Borders, Ingram Book Co., and the Online Computer Library Center.

You’ll also use these contacts when creating metadata files for your newest projects.

Remember that although you may see metadata referred to as ONIX—ONline Information eXchange—this is only a format, an XML (extensible markup language) DTD (document type definition). Most organizations within the book distribution channel do not require submissions in ONIX format.

If you are dealing with firms that do require or expect data in ONIX, or if you would prefer to outsource the dissemination of your metadata, you can work with a vendor such as:

Firebrand Technologies

Fran Toolan describes his company by saying, “We offer software that helps publishers track projects from acquisition through publication.” Its Eloquence Metadata Solutions service distributes bibliographic information to such retailers, wholesalers, and aggregators as Amazon, Barnes & Noble, Ingram, Baker & Taylor, and Bowker. Monthly subscription rates are based on the number of active titles.

HeadStart Software, booksonix.info. Customers of this firm can upload metadata into their own databases on the Booksonix server. Booksonix then disseminates the information in whatever format the recipients prefer. Subscription fees are based on the number of titles.

NetRead Software and Services, netread.com. The company’s Web site describes it and its metadata product as providing “marketing solutions that publishers access through their web browsers.” Its JacketCaster is an ONIX-conversion service available for as little as $99 annually for publishers with 25 or fewer ISBNs.

More Motivation

Want another reason to ensure you have a correctly formatted 13-digit ISBN printed on your book? Cost. At Barnes & Noble, for example, you’ll pay 13 cents for each book that must be stickered with an ISBN-13 with the price embedded, and with a human-readable price in the correct size and location, in the correct colors.

Part 3:

Blame It on the Nature of Publishing Today

If you feel as if the “Desperately Seeking Good Data” series has been bombarding you with demands to get your book data together and get it to Bowker, distributors, wholesalers, and retailers accurately, completely, and much earlier than you think is possible, you may find Scott Lubeck’s perspective on metadata refreshing and reassuring.

Lubeck, the recently appointed executive director of the Book Industry Study Group, which handles metadata best practices and standards for the United States, can calmly explain how these dozens of pieces of information and the way they’re submitted came to be so important and how publishing is only one of many industries dealing with metadata today.

When we talked, he started out with a reminder: “Metadata is an intimidating word. What we’re really talking about is, very simply, how to describe a book.”

This is nothing new. Customers have always wanted to have books described to them. But when distribution was less complicated, someone other than publishers took care of most of this description—and less of it was required.

As I mentioned earlier in this series, today my local bookseller has 144 references to Jane Austen’s Emma in her online database. I don’t know how many listings there were 10 or 20 years ago, but I can almost guarantee you that there were far fewer—perhaps as few as 25 percent of today’s references.

When I started publishing books in 1990, booksellers were relying on microfiche from large wholesalers, paper catalogs from smaller wholesalers and publishers, and Books in Print—in print. If a customer wanted to know which edition of Emma to order, a bookseller, often the store owner, may well have been the one who made a recommendation, based on having seen display copies. (Remember, until 15 or 20 years ago, lots of communities had independent bookstores; and in 1989, Barnes & Noble had only 23 stores.)

“We can’t be so casual today, because information is exchanged on a global basis and exchanged mostly online,” says Lubeck. And, of course, much of this information is accessed without a bookseller—or any other human.

Moreover, what he calls “discoverability”—the ability to find what you need and purchase it online—is now important in every industry: “This demand for really good information comes from the consumer.”

What Matters Most

But in case you’re still feeling overwhelmed, he emphasizes that metadata is not a technical issue: “It’s an analytical issue, like an index.” What BISG wants publishers to recognize is that the quality of the information is far more important than the ONIX tags.

Lubeck cites two kinds of stumbling blocks for publishers as they strive to provide good descriptive data about their books. Some publishers think of themselves as craftspeople. Many of us started as writers and continue in the dual role of writer/publisher, and we are impatient with chores such as inputting metadata. Other publishers, perhaps especially those in large companies, don’t realize that when roles are defined in terms of departments and books are passed from one department to another, nobody is taking responsibility for ensuring that all data about them is correct.

Although some firms use title management systems that force the gathering of metadata, Lubeck says publishers can operate with a simple spreadsheet—as long as specific people are accountable for the information on that spreadsheet.

In general, Lubeck says (and others in the industry interviewed for this series of articles agree), almost no publishers provide data six months in advance of on-sale dates, and most publishers are unaware that much of the data they enter is inaccurate. Ask publishing executives to evaluate the quality of their metadata, he notes, and most give themselves high marks, failing to recognize errors and gaps.

How to Ruin Your Reputation

But errors and gaps in book information can create serious credibility problems for a publisher with everyone in the supply chain. “Bad metadata creates a bad impression of the publisher in the same way that a book riddled with spelling and grammatical errors does,” Lubeck explains. “Being ‘almost right’ isn’t good enough: it tells the buyer not to trust you.”

A publisher often has only one chance to sell to a customer, he emphasizes, and if that customer—often a bookseller—commits funds for a book that does not arrive as the metadata says it will, or if it is not what the metadata says it will be, it’s unlikely that a good long-term relationship will develop. A publisher can expect a similar problem with its authors if its metadata has errors, especially if the errors are not corrected, Lubeck believes. A publisher, he notes, is a steward of metadata, and he suggests asking yourself, “How good a steward are you?”

Besides working to help publishers recognize the importance of metadata—accurate, timely metadata—the BISG executive director says his organization wants to see technology that is affordable for even the smallest publisher, so that almost all data can be submitted in machine-readable form, even by one-book publishers with day jobs. A U.S. program similar to the BookNet Canada’s BiblioShare would be “great,” Lubeck says. (For more information, see booknetcanada.ca.)

Correcting Errors: A Cautionary Story

Carin Smith of Smith Veterinary Consulting and Publishing Services was checking on the class-action suit against Google last winter when she discovered an error in her book data on Google’s Web site and—even worse—on Amazon.com. Because her metadata is correct on Bowker’s site, Smith suspects that many online retailers are using Amazon, rather than Bowker, as their source.

“It’s frightening how an error can be magnified when so many retailers use a single other site for their information, especially when they are using another commercial site,” Smith says.

To complicate the situation, these errors have not been easy to correct. Smith contacted Amazon.com, which responded by deleting her company’s name from the book description and retaining the wrong publisher’s name. She also notified Google—but three months after she sent her email, the inaccuracies persisted. (And she adds that Google told her they will not be corrected until the Book Rights Registry outlined in the Google Settlement is established. Updates about the settlement and its progress will reportedly be posted online at googlebooksettlement.com.)

Smith’s advice for other IBPA members:

* Check every detail of all your listings on Amazon.com and on the Google settlement site.

* Check your listings after you’ve been notified that errors have been corrected, and see if your information is truly accurate.

* Enter the erroneous information in a search engine to see how many online retailers and other Web sites copied the inaccurate information into their text, and notify them of the errors. (By entering the errors in Google Alerts, google.com/alerts, you will be notified of those Web sites that continue to use bad data.)

Linda Carlson (lindacarlson.com) writes from Seattle. She spent more than 10 years publishing job-search guides but had never spoken in terms of book metadata until beginning this series of articles.