Demand modeling
for Genre Fiction Authors
Nat 'Nose' Connors
<nat@kindletrends.com>
Book In a Month June 2026
This talk is going to be a bit more nerdy than usual
although not in a tech sense
There will be practical techniques
but there will also be some abstract waffly big-brain stuff
Me (v. briefly)
A technical writer
IT project management
Cancer research
...now, mostly, I just write about lurve
Kindletrends
-
Weekly and monthly information for your genre
- First month free
- You can cancel anytime you want
- One price for everything. No upselling, premium tiers, ads, or sponsorship.
- Use the code BIAM, and subscribe for USD10/month forever.
Some free resources:
Listing all Kindle Store categories
Getting Also Boughts
Searching the Kindle Store
Follow along!
This is a web-based slide presentation. To go forward, press the space bar, use the arrow keys, or swipe if you're on a touchscreen. Press 'O' to get a view of all the slides.
Assumptions
-
There is such a thing as 'unmet demand' for books, and it has a shape.
-
Customer behaviour has some relationship to this demand. So if a customer somehow indicates that they want <x> in a book, they are more likely to actually buy a book with <x> in it.
-
The data we can access is enough to make useful models of this demand.
-
This demand is actually a lot of separate little things which are quite specific to groups of readers, and are constantly changing over time.
But
-
As creators we also play a role in shaping this demand with our work, so in truth there's a feedback loop.
-
There are a bunch of other reasons for understanding the market, to do with creative fulfilment and our own personal situations. Right now we're just focusing on the dispassionate 'find out what readers will buy and then sell it' aspect.
-
There are a lot of different ways customers can 'indicate' things, and some of them are a lot more useful than others.
And an assertion
-
Books are not (very) 'fungible'
-
Books are not (very) exchangeable
-
Our emotional relationship to our products is complex
-
Our readers' emotional relationship to our products is complex
This job is not selling widgets
=> be cautious of typical sales talk
So what are we doing?
Right now we are mostly only getting a picture of supply rather than demand.
We do this by looking at books published on Amazon and other platforms.
Why are we looking at supply?
-
There's a lot of it -
It's relatively easy to get -
It's relatively easy to translate into action
(e.g. if a lot of authors are using illustrated covers in your genre, maybe you should be using an illustrated cover)
But demand is a bit different
-
Google Trends search volume
-
Scraping Amazon autocomplete
-
Whatever MerchantWords does
The problem with these measures of demand
-
They aren't very specific
-
They generally won't help you adjust course on your existing product
-
They are all 'pull'-oriented: you need to know what you're searching for
So where could we get more specific demand information?
Other sources
-
FB group discussion
-
Discord group discussion
-
Tiktok
-
Amazon carousel information
But what's the point of all this?
-
Identify unmet demand ahead of time, so we can align our books with it
-
For a given author/book/brand, profile its relationship to demand
& what can we do now vs the future?
So where are we at?
Step 1: Categories
This (probably) won't work
In principle this ought to be easy
1. Point Claude at this:
2. Prompt it 'identify categories with high rank and low activity'
...
3. Profit!
Beware
- Books can be in more than one category
- Categories can be driven by only one author, or a small range of authors
- Categories can be 'high sales, low supply' because of what's in them - not the other way around
- This doesn't tell us much about the 'why' of people's buying patterns
- At the category level, these measures are actually pretty static
So what use is this?
- Understanding the 'neighborhood' of categories surrounding a particular concept or genre
- Orienting yourself to terminology
- Seeing the breadth of approaches in a particular field
Saturation of demand
Lucalia:
We look at data to see what’s selling, but does Kindletrends help us see where the dead zones are? Specifically, can it help us identify categories that have high reader demand but a sudden drop-off in fresh content?
Market awareness is great, but by the time a trend shows up on a newsletter, it’s often already peaking. How can we use Kindletrends data to differentiate between a short-lived micro-trend and a sustained, viable trope that’s worth spending six months writing?
Melody:
What are some ways we can try to spot trends before they're peaking and oversaturated? The thing is
What can we do practically?
Find the work that has longevity
...and understand its neighborhood
Weeks by book
Weeks by author
Find the work that has longevity
The point
...and understand its neighborhood
Step 2: Also Boughts/Also Reads
Why?
-
Also Boughts/Also Reads represent concrete reader behaviour, giving up money or time
-
They are a buying signal - so not just about the cover/name
-
Look for situations where a midlist author has a lot of 'cross-referenced' Also Boughts; they are doing something right
Getting a reader to buy the first book is a matter of communication
Getting them to buy the second book is a matter of content
AI 'slop'
Lucalia:
With the influx of high-volume, AI-assisted content flooding specific sub-genres, are you seeing these become artificially crowded? How can human authors use Kindletrends data to spot and avoid niches that are currently being targeted by mass-produced content bots?
Reimers & Waldfogel (2026)
"AI and the Quantity and Quality of Creative Products: Have LLMs Boosted Creation of Valuable Books?"
The response was about what you'd expect
But I went and read the paper
Counting books per month
"We obtain the number of books published… using queries to Amazon's advanced book search function at amazon.com/advanced-search/books. We request new, English-language ebooks, and we specify their publication dates by month and by category for all 30 Amazon categories… A query delivers a list whose first page indicates '1-16 of x results.' When there are fewer than 1,000 results, x is the exact number… Up to 10,000, … nearest thousand. From 10,000 on, … nearest 10,000."Counting books per month
"Seven of the 30 categories...become sufficiently large during the sample period that the monthly category totals are reported only to the nearest 10,000. For those categories, we obtain their 148 subcategory totals by month to create the aggregate time series of new books released by category."But books can be in more than one category(!)
Counting the whole cohort suggests current rates are about half what is suggested here
So I think this paper is double- or triple-counting some books
So what?
A better view
The point (1)
-
I think there's evidence publication rates have increased 2-3x since 2022 (across the whole store)
- I don't think there's evidence there are 300K books being published per month
- FWIW their AI detector (Pangram) suggests that only ~20% of post-2022 fiction books sampled are 'AI' (with numerous caveats)
- Some genres (e.g. Sports Romance) have grown hugely while not reporting high levels of detectable AI use
The point (2)
-
If you are writing nonfiction you may have a problem with low-quality machine-generated content
-
D2D now has a subject matter expertise requirement
-
If you are writing fiction, you may occasionally see niches where there is an influx of low-quality machine-generated content
-
But at the genre level the challenge is mostly the same as it always was: getting people to buy the second book
Questions
Indie vs trad
Melody: What are some ways that applying market trends works differently for indie authors as opposed to traditional publishing? Are there some strategies that tend to work well for big publishers, but not for indie authors, or vice versa?
Categories & keywords
Melody:
Since Amazon has changed the limit from 7 to 3 categories, what kinds of changes have you seen in the markets per category? Sometimes it feels like it's no longer worth it for an indie author to try to rank high in a category, as we're no longer playing the game of "try to rank in a lower-performing category so you can get a #1 new release banner and get more visibility upon release." Would you say this is accurate, and that our efforts would be better spent focusing on other strategies such as keywords?
Nicole:
When, why, and how do you update your keywords for older books, if at all? I can't update some of mine due to content and risk of being banned, and I know many authors do this every 6 months. Is there a trend or anything in particular you look for instead to make it worthwhile and have the best chance of impact?Backlist
Melody:
What kinds of trends do you see with new releases vs older books? Is it possible for a book from the back list to get as much visibility as a new release? Do you see new releases funneling readers into an author's back list enough to make older books trend again?
Finally
- I think we are actually modelling demand, despite how it may seem
- But the most effective ways to do it (right now) are still pretty subjective
- And they are at the level of style and emotional impact rather than baldly at the level of 'content'
So, the summary
When you are researching,
focus on the shared emotional experience that is delivered by successful work
And
the attempt to replace judgment by formula is always irrational;
all that can be done is to make judgment possible by narrowing its range and the available alternatives, giving it clear focus, a sound foundation in facts
and reliable measurement of the effects and validities of actions and decisions
Kindletrends
A weekly and monthly market research newsletter
Research done for you, in your inbox every week
Summaries
Interactive tools for research
Alerts about books
Cover montages
Blurb trends and details
Filter for specific elements
14 genres in one subscription
Romance One-Hour Romance (KSR) Paranormal Romance Gay Romance Historical Romance Science Fiction Romance Mystery, Thriller & Suspense Cozy Mystery Science Fiction Fantasy Teen & Young Adult Urban Fantasy
-
Your first month is completely free You can cancel any time you want
Use the code BIAM for the next week, and subscribe for USD10/month forever.
Acknowledgements
Lana Love
Becca Syme
AJ Lancaster
TK Eldridge
Lizzie Dunlap from Pixie Covers
Elizabeth Brady
Thank you to all the authors and artists who helped with this talk
Nicole for giving me a chance to speak
All of you for your time and attention
Thank you for watching!
For more free resources, check out:
For the slides for this talk: