Categories and keywords on online retailers: choose them wisely and the algorithms will target your ideal readers – especially on Kindle. You can make a whole science out of it, but this piece on KDP explains the basics in good, plain English.
Essentially, you pick two categories, and then get yourself in several more specialised lists by including set keywords.
But this system has its limitations. At first writers of genre fiction had many sub-categories to choose from, but writers of literary, contemporary and general fiction found themselves in one immense category where it was hard to be seen. There were few ways to tell the algorithms ‘I’m non-genre but I have a flavour of romance, or loss, or my novel is set in Borneo’. Recently Amazon has made big improvements and refined the choices – find them here.
Despite this very welcome addition, the results haven’t been as good for me as when I unknowingly broke the rules. When I put other authors in the keywords, my sales soared.
Tsk tsk
I did it in all innocence. Reviewers had been comparing my first novel with Paulo Coelho, Margaret Atwood, John Fowles, Doris Lessing, so I put those names in the keywords. My sales rose, readers seemed happy to have found me this way – so the comparisons must have been useful and valid. Then I discovered writers who did this were being sent warning emails so I removed them – and fizzled back down the charts.
It’s a real shame, because for me, this tactic was more effective than keywords about genres, subjects, settings, themes and issues. And surely the author and their style is a significant feature of any novel. With literary fiction, it’s the most important quality of all. It’s a valid way to talk about a book in the literary world – and yet it isn’t accommodated in the search mechanisms that writers can control. It’s a refinement that would be helpful to both authors and readers.
What’s more, now would be a great time to discuss and lobby for it. Here’s why.
We are connected…
Last week I was watching a videocast from the Grub St Writers Muse and the Marketplace conference. One of the panel members was Jon Fine, director of author and publisher relations at Amazon, so I tweeted @Grubwriters with my point about author comparisons. Jon Fine was rather interested in the idea and replied that it was something they’d never thought of. So…. watch this space!
(Let’s pause for a geek check: I tweeted a question in my home in London at 7.30pm, watched it read out to a room in Boston where it was 2.30pm, and real live people started to talk about it, with voices and hand-waving… and a man from Amazon stroked his chin and said ‘maybe we could…’)
So I want to kick off a discussion here. Amazon are in the mood to get constructive feedback on this right now. There couldn’t be a better time to discuss it. I’ve shared my one tiny idea for improving the algorithms to help readers find our work; you guys no doubt have more to add. The questions begin!
1 Have you tried a category tweak that got you to more readers – Amazon-legal or not? Is there a category facility you’d like to see?
Jon Fine also said the categories problem was more widespread than Amazon. The industry standard for classifying books by subject, BISAC seems limited in its precision, although possibly it’s geared for booksellers rather than readers.
2 If you are – or have been – a bookseller, what’s your take? Would you find it helpful if the BISAC categories were made more flexible and detailed?
3 As a reader, how do you use search tools to find new books?
Let’s discuss! And change the world… 🙂
Nail Your Novel 
Thanks Roz! This is cool.
Your link to the Amazon category list wasn’t working so I had a dig around – did you mean this one? https://kdp.amazon.com/help?topicId=A200PDGPEIQX41 – A treasure trove indeed. I’ll be going through and tweaking keywords on ALL my books!
Thanks, Grace! I could have sworn that link worked when I first wrote the post, but I’ve grafted yours in instead.
I’m guessing that this is pretty common when targeting ads for goods and services, so why should books be any different? I was looking for a particular supplier yesterday and found that by typing part of their name into a search engine I got a list of people who supplied the same product, something that was really useful.
My only reservation is that some people will add tags to their books that aren’t accurate in order to increases sales. Perhaps they could restrict similarity to matters of fact, so you could say your book was ‘like’ another if it dealt with the same subject or was set in the same location, but not, for example if you felt it was written in the same style.
Alexander, you’re right that the comparison could easily be abused. But all the categories could be. The only way to police it would be by the responses of reviewers. If readers were dissatisfied with the comparisons I’m sure they’d make their feelings known!
All of the vendor portals I use offer a limited selection of BISAC categories when you set up your book, but as you pointed out, their sites offer readers a lot more “custom” categories to choose from. Since the authors don’t get to tell them what custom categories are appropriate, the vendors have to bridge the gap using something else. Enter keywords. For my books, I enter Swords & Sorcery as a keyword so Amazon places me in that category. Works great, even if it is an odd way to solve the problem.
I understand the rationale from the vendor’s perspective. They offer BISAC categories because that is a standard that changes very little over time. BISAC categories flow fairly reliably as metadata between distribution partners. The vendors don’t offer a selection of their own custom categories because (at least in Amazon’s case), they tweak them occasionally, and they would NOT flow reliably between distribution partners. Instead, vendors like Amazon let authors “hint” at what custom categories are appropriate through keywords. As long as the keywords are an accurate reflection of the content of the book, that theory works reasonably well. In my case, even if the Swords & Sorcery category were to be eliminated, that keyword is still an accurate reflection of my book’s content.
I don’t think there’s an easy solution to the categorization problem. The best solution is to have the standard (BISAC) categories reflect how readers want to find books, but that is definitely a moving target. A standard that changes loses cohesion as a standard unless everyone keeps up with it and modifies their sites to match it. I think Amazon is doing about as good as it can right now.
That said, the idea of offering some kind of style hint is an intriguing idea. I’m sure readers would love to be able to find more authors who write like their favorites. Unfortunately, that feature would be instantly abused by the over-zealous authors who would try to associate their name with the hot author of the moment. Maybe what we need is an *accurate* I Write Like (iwl.me). I’m not holding my breath on that one.
Hi Daniel – all good points. And you’ve brought up the problem of abuse, with a category that’s subjective. I guess this is the point where algorithms rub up against gut instincts and humanity, and are trying to replicate what goes on in a reader’s heart and head.
Reblogged this on Charlotte Gerber.
Thanks, Charlotte!
I believe Apple and B&N allow five categories each. Why doesn’t Amazon? And why did they take the tags away? Would like to know the answers to these questions.
Jim, I never really understood how we were meant to use tags. I seem to remember they disappeared about the time the improvements were made to categories, so perhaps Amazon think it’s a step forward. Have you noticed a disadvantage now we don’t have tags?
I’d love to see more categories & keywords. Smashwords lets you choose as many categories as you like.
I write mixed-genre books and they’re all different. It’s a marketing nightmare. I want to make it clear to readers what to expect from one of my novels and I know from experience the easiest way would be to mention authors’ names (and this is what bloggers & readers do in reviews). I want to refer to a certain type of accessible literary-ish fiction, but since this is frowned upon I confine these refs to my blurb.
For my latest novel CAULDSTANE (literary/paranormal/romance/psychological) I selected 7 keywords: “gothic love story, ghost story, contemporary romance, Scottish fiction, older heroine, paranormal, literary fiction”, but I would have liked to have mentioned Daphne du Maurier & Mary Stewart. Those authors’ names would have taken me straight to the readership I think would enjoy CAULDSTANE, because the book owes something to the novels of those authors (literary idols of mine to whom I acknowledge a debt.)
I don’t think this is cashing in on another author’s success. It’s helping readers find the right book.
But if Amazon can’t/won’t do this, could we PLEASE have more sub-categories for literary fiction? It’s the Cinderella genre.
Hi Linda! It’s funny that Smashwords allows so many, and we hardly see any articles about making the most of Smashwords algorithms. Perhaps Smashwords doesn’t have so many recommendations engines, or perhaps we’re missing something we should take more notice of.
You’re saying exactly what I feel about author comparisons – sometimes they are the only reliable shorthand to convey the experience of a book. As you say, it’s not dishonest or misrepresenting. At least when I did it, no one seemed to think so!
More sub-categories would definitely be welcome. When I look down the list, although there are many more choices, they’re not useful for my books.
It’s pretty easy to get into categories by putting those category names in as keywords. Whether this increases visibility and discoverability, I don’t know how much. My books have been compared to early Kurt Vonnegut. If others compare your writing to someone else, surely there must be a way to tap into this? How about a compared to category, all legal and above board.
Hi Eliza! A ‘compared to’ category would be a brilliant solution. Then it’s clear. Brilliant suggestion.
I was thinking the same thing – let the crowd vote (perhaps allow the author/publisher to place up to 3 ‘similar to’, and then let reviewers vote on those and their own submissions). Thinking about it, this is really just a reintroduction of tags, but they would only be able to draw from the author name database on Amazon. I think it would be important for the readers to remove an ‘is like’ too.
I do think there is a need – as in a bookstore, where the bookseller would quite likely know ‘this author is a bit like’ [I’ve had plenty of personal experience of this!] – and not just to dump in a genre. There’s a big difference between George RR Martin (who I cannot read) and early Terry Goodkind, but they are both classed as ‘Fantasy’.
Perhaps the ‘Readers compared Roz Morris to’ section could actually be a decent algorithm and a graph (the same way that Amazon use the reviews, eg: “53 other reviewers made a similar comparison”
Good refinements, Robert. The voting system might make it more accountable and reliable.
Netflix has about 600 categories for film. And as a result, they’re really good at recommending movies I might want to watch. Amazon could do the same for books. It’s just programming. Surely they could figure it out. http://www.allthefixes.com/listing-of-all-netflix-com-movie-genres/
Joe, that’s a useful comparison. There must be a lot of crossover there tha Amazon could use.
I’m not there yet, but this is really useful. Thanks Ross.
My pleasure, Ashen!
While the tweaks that Amazon have been making to categories are a step in the right direction (by permitting the use of keywords to place a book in a sub-category, or creating a situation where two categories–like mystery and romance help create additional categories that combine the two Romance–mystery, etc.) I really think that one of the most helpful changes would be to go back to the minimum of 5 category choices.
I know for a fact that one of the reasons my first book sold better than my second was that the first was published back when I could chose 5, and these got grandfathered in, while the second only got two.
I know that specific authors can get more categories (as I have), and I just noticed that a newly published book I looked at had a grand total of 12! But I also know lots of authors who by the nature of their main categories just can’t get more than 2, and that really hampers their ability to reach their markets through different strategies.
In addition, to complicate matters, the book browsing categories are not identical to the ebook ones, and the US browsing categories aren’t identical to the UK ones, etc. I understand that this is probably because computer codes were set up early on that take time and money to modify every time there is an expansion of categories. But it would be nice if we knew at least that their goal was more standardization across stores, and increased ability to have your book in more than 2-3 categories.
Five categories? Wow, I don’t remember that and I thought I’d been doing this a while!
It certainly is confusing that the categories are different between ebooks and print. I think Amazon has to be hampered by BISAC for print, instead of its own finer-tuned system.
And I’ve heard that books from traditional publishers seem to get more categories – which may be why you saw a book with 12. That clearly isn’t a level playing field!
Thanks for such a detailed comment.
We put comparables in our query letters to editors so I see nothing wrong with listing comparables in marketing our books. Since we are all unique writers, giving a mix of comparative authors doesn’t seem like trying to hang on to someone’s coat tails.
Thanks, Joan! It’s funny how the algorithm-makers haven’t tumbled to this, as so many people here seem to think it’s an essential part of choosing a book. Perhaps the Amazon folk have thought a little too much like machines…
Thanks for diving in and opening up an avenue of conversation Roz!
Allowing authors to indicate a book is a ‘combination of’ certain titles or authors might also be helpful, considering indie’s are doing a lot of non-traditional crossing of genres and tropes…
I say specific titles, because sometimes an author’s work is really varied: for example your book may be like a specific Michael Crichton book (eg. Jurassic Park) rather than say his novel, Timeline.
Good point about the book comparison, Emily. And as you say, indies are more likely to depart from industry standards. That’s often the very reason why some of us end up as indies at all.
Three of our last five books have been anthologies, so author comparisons are baked into the mix for us. With one of our anthologies, comprising the audience favorites from a popular women’s poetry festival, we were conflicted about which two Amazon categories to use. “Women’s Poetry” was easy, but while either “American Poetry” or “Contemporary Poetry” seemed accurate, we ticked off “Gay & Lesbian Poetry” for our second category.
We’ve been #1 in /that/ category, excepting only a few weeks, for seven months.
Mind you, several of our contributors are lesbians, and several poems in the anthology deal with LGBT issues. But virtually ALL our contributors and subject matters are Americans, and ALL of them are contemporary. Did we paint with too broad a brush? Given the nature of the festival that was our source for the book, it’s unlikely any of our contributors would feel /offended/ at being lumped in as LGBT poets, but such a label was (is) hardly accurate. However, we felt, and still feel, that the choice of our two categories made sense /for the readership we imagined would seek out and read our book/.
The chart-topping success, for better or for worse, tends to paper over some of our misgivings.