Home » Discovery » Why publishing needs tagging

Why publishing needs tagging

21 November 2015

From boingboing:

Walk into a bookstore, and chances are you’ll see books divided into sections by genre. Romance, Science Fiction/Fantasy, Mystery, and so on. It’s the most common system of categorizing books, conversationally and from the data-management perspective of the book world. Genre is also incredibly limiting at times.

There are dozens upon dozens of subgenres across the genres of popular fiction (Romance, Crime, and Science Fiction/Fantasy, plus some others). Science Fiction gets sliced up into Space Opera, Mundane SF, Hard SF, Cyberpunk, Dieselpunk, etc. These subgenres can get hard to keep track of, especially since their boundaries are often porous, and even life-long fans often disagree on the borders between sub-genres, policing them inefficiently but with gusto. At times it’s fun to argue classifications, try to find exactly the right place to frame a piece so that its cultural and narrative context is most clear. And narrow sub-genres can be useful for putting works into clusters for conversation, but it’s also really easy to slice so thin that the discussion becomes obscure or self-serving rather than practical.

Ultimately, a hardline This-or-That, pigeonholing system of defining genre and works is far more trouble than it’s worth, and can do a great disservice to works that defy easy categorization. Most traditional systems in the publishing industry fit into the pigeonholing system. BISAC Codes (basically trade publishing’s official genre system) are fairly granular, but totally fail to keep up with the proliferation of sub-genres, and the genre categorization systems of all of the major ebook retailers fall victim to the same This-or-That approach.

But there is hope. And unsurprisingly, it comes from the internet.

The Tag. You know, this little thing: #

Whether it’s hashtags on Tumblr, metadata tagging in LibraryThing, or conversational hashtags on Twitter, readers conversant in internet discourse are more and more familiar with the concept of tagging works – using a Yes-And approach for categorization instead of This-or-That.

Tagging is incredibly liberating. A work can be tagged #urban fantasy, #YA, #LGBTQ, #domestic violence #Cleveland #1970s, meaning that if a reader has access to those tags as part of the information given to them about a work, they can very quickly identify its major facets.

. . . .

Let’s take some well-known titles and re-present them with a set of tags.

Harry Potter and the Sorcerer’s Stone (aka Harry Potter and the the Philosopher’s Stone) #Fantasy #Middle Grade, #Magical School, #Friendship #Prophecy #Rags to Riches #England #Hidden Magical World #Series

Or The Martian, by Andy Weir #Science Fiction #Adult #Mars #Survival #MacGyver #Epistolary #Space Exploration #Thriller #Disco #Man Vs Nature #Stand-alone #Voice-Driven #Try-Fail Cycles

And looking back at a classic, how about Pride and Prejudice, by Jane Austen: #Romance #Adult #Regency #Hate At First Sight #Suitors #Order of Marriage #Sisters #England #Gentry #Banter #Drawing Room Politics #Gender

Link to the rest at boingboing and thanks to Meryl for the tip.


14 Comments to “Why publishing needs tagging”

  1. You mean…

    Harry Potter and the Sorcerer’s Stone,
    Books, Children’s Books, Growing Up & Facts of Life, Friendship, Social Skills & School Life, Friendship, Science Fiction & Fantasy, Fantasy & Magic, Action & Adventure

    This kind of labels?

    Oh. Did anyone tell Bezos?

    Take care

  2. I was just in a genre discussion on another forum. I think the question was about being forced to write in a box and I said it was about finding the correct box for what you had written.

    Tagging would certainly assist with cross-genre work. I go to Amazon and look at all the additional tags people have invented for their books. Sometimes I’m surprised by a new subgenre that fits my stuff.

    But they are all hashtags, so where does one apply these? That’s way too many for a tweet.

  3. There is no Fiction>Sports>Equestrian in BISAC.
    That’s not drilling down very far.
    Amazon doesn’t care.

    • There OUGHT to be.

      There’s a HUGE market for grown up horse stories. All the women who read horse stories as girls could be interested in:

      Equestrian/Zombie Apocalypse

  4. Much of Harry Potter takes place in Scotland. Just sayin’ 😉

  5. Amazon had tags: anyone could add to them, and pretty soon everyone did. And many of them were rude, crude, and socially unacceptable, basically attacks made anonymously. So now we don’t have tags.

  6. This piece is entirely well-meaning but completely wrong-headed:

    1. This problem doesn’t really exist – it’s already more-or-less solved by keywords and categories (and on another level by things like genre convention in cover design, or, indeed, the book description, title or sub-title, and even very occasionally by the publisher/imprint name).

    2. Many retailers (e.g. Amazon and B&N) don’t actually use the BISAC category list – presumably because it isn’t comprehensive, diverse, or granular enough. They have a plethora of additional categories and sub-categories, with country- and language-specific categories. It’s extremely comprehensive and they add to it all the time – again, presumably based on customer/supplier demand. This is one area where we don’t have to deal with tradpub intransigence/inertia/technophobia.

    3. The existing system isn’t This-or-That. Far from it. On Amazon you can choose two categories when uploading. You can choose three (or more) on the other retailers. And, on Amazon, you can add additional categories through selecting certain keywords. To take his bookstore example, it means your title can appear on multiple shelves simultaneously.

    4. The proposed solution doesn’t work (I). We know this because… we already have it and it doesn’t work. Not all retailers allow authors/publishers to choose keywords. In such circumstances, authors/publishers often resort to appending the keywords at the end of the blurb as a crude workaround (I actually wonder if this is why the Angry Robot store does it – i.e. I wonder if it can properly handle keywords). If you ask anyone who uses such a system, they’ll tell you it’s not superior to being able to choose targeted keywords that will properly trigger appearance in search – aka the system we currently have on most retailers, such as B&N and Amazon.

    5. The proposed solution doesn’t work (II). This would be a fail on the customer side. Take his suggested examples and tags. How is tag-searching on “England” or “Disco” or “Series” going to be useful? Such categories could include all kinds of fiction or non-fiction and are thus far too broad to work for either publishers/authors, customers, or, indeed, the retailer.

    6. The proposed solution doesn’t work (III). We know this because… we already had it. Amazon had a tagging system until (I think) 2011 and it was ditched because (presumably), as an open and unregulated system it was abused by authors/publishers and readers, as an unlimited system it resulted in tag spam and was, thus, not a useful way to discover or categorize books, and it attempted to replicate the work already being done much more efficiently by the ever-expanding category system, and the keyword+search system.

    7. The proposed solution doesn’t work (IV). Let’s take a look at the model for this proposed system – Angry Robot’s store, The Robot Trading Company. Now, I’ve been an Angry Robot fan since before they launched the store back in 2012. And it’s not at all how you would expect a publisher-run store to look. They are future forward, progressive, not afraid of technology or the internet, etc. Books are cheap and DRM-free – allowing them to sell Kindle-compatible versions – they have a clear brand, good authors, strong covers, and titles are generally marketed with quite a bit of savvy and good humor. That’s the good part. The tagging they do is not an innovation I would highlight, let alone want to see replicated. It’s kind of fun and sometimes funny, but it’s not very useful. Taking a random example from their homepage – The Ark by Patrick S. Tomlinson – I looked at these tagwords (totally taking credit for that invention) and then searched based on them to find something similar. I chose the “We’re Not Alone” tag. Ignoring the laboriousness of cutting-and-pasting, the results are as useless as you would expect. And right away I can see the sequel to a book I read last week (Apex by Razem Naam) which has no aliens/extra-terrestial elements. Or tagwords. Oops. In fact, there is an even more basic metadata fail here. The title or author aren’t clickable. I’ve no way of quickly seeing what else the author has written. This is the third in a trilogy, where are the other books? I can’t find them from a simple search. Do they sell them? I don’t know, but I do know I’ll shop somewhere else next time because they don’t seem to have the basics down. Maybe they need to fix those before attempting to solve problems which don’t really exist, and which are already kinda solved.

    A deeper issue: the whole piece is based on a flawed assumption: that we have a Discovery Problem. If someone wants to posit that readers are having trouble finding books (or good books, or quality books, or books they like, or any kind of books at all) then they should bring some evidence to the table because I just can’t see how that is possible.

    I think the author meant to suggest we have a Discoverability Problem – which is a different thing altogether, a publisher/author problem, rather than a reader problem, meaning, in short, that people aren’t discovering *your* book… and reading someone else instead. To solve that problem, well, it might be a good idea to bone up on keywords and categories. First lesson: they exist!

    • Thanks for the in depth response. It confirms some of the things I thought I knew.

    • I would only caution that tags are not meant to be searched individually. Of course they provide no value in that context. It takes keyword matching on 4 or five points of data to create a proper corrolation. So, if you’re looking for something for fans of Harry Potter, you don’t search for “England”, you search for “England, wizards, youth, school” within the fantasy genre.

      Simply clicking on a single tag would be a rather stupendous waste of time, I agree.

  7. Al the Great and Powerful

    So I’m stuck sorting with keywords some random person made up, trying to guess what they meant, and how it will meld to fit my search?

    Oh HELL no.

    I’ll use an industry list, or an Amazon list, but I am not going to support general tagging as any kind of acceptable.

  8. Uh…librarians have been doing this for nearly fifty years. It’s called MARC (MAchine Readable Cataloging), and it goes way beyond fiction.

    The whole point isn’t to search for ONE tag — it’s to search for many tags at once. Give me five titles you like, I’ll find the tags all five have in common, and search for other books that have the same combination of tags.

    The biggest problem is time…there’s usually not enough time for any cataloger or cataloging system to enter more than a dozen or so of these type of tags.

    A few years ago there was an outfit that used supercomputers to do textual analysis and apply weighted scores to tags for each title, i.e. 73% vampire, 86% romance, 3% horror, 45% young woman, 22% Lesbian, 63% New York City, and so forth across thousands of terms. They used this info to analyze books a reader liked and find similar books across several dimensions. They were phenomenally successful in matching readers & books.

    Amazon bought them.

Sorry, the comment form is closed at this time.