Profitable Fiction KDP Niches in 2026: I Tested 4 Genres. Only Cozy Mystery Was Indie-Friendly.

I'm a first-time KDP author from Atlanta. A few weeks ago I tested 25 non-fiction KDP niches with live Amazon data using my own tool, BookRankPro. The standout opportunity in non-fiction was menopause survival guides.

Then a reader asked me to research the niche around Theo of Golden, a #1 NYT bestselling literary novel. That's fiction. My tool was built for non-fiction, and the data it returned for fiction was misleading. So I rebuilt the relevant parts and ran the same live-data test on four fiction sub-genres.

This article is about what changed and what I found.

Total niches tested: 4. Total credits spent: 11. Total surprises: more than I expected.

Three of the four niches were dead ends for indie KDP authors. The fourth had real opportunity, with a heavy caveat about publishing volume. The bigger story is what I learned about why fiction niche research needs different methodology than non-fiction.

TL;DR (the 5 things that actually matter)

If you stop reading here, take these:

Generic fiction terms return zero books on Amazon. "Small town novel" came back with no results because Amazon collapses broad terms to ads or non-book products. Fiction research requires specific sub-genres and tropes.
Earnings math breaks for fiction without per-book median. Averaging BSRs first and converting once produces nonsense numbers because the BSR-to-sales curve is non-linear and fiction has high BSR variance. Per-book earnings is the correct approach.
Publisher matters more than reviews in fiction. A niche where 3 of the top 5 books are Big 5 trad imprints is uncompetable for indies regardless of what the earnings or reviews look like. Reviews alone are a misleading signal.
Kindle indie books have no Publisher field at all. Amazon omits the field because the author is the publisher. The absence of a publisher in a Kindle PDP is itself a strong indie signal once you know to look for it.
Of the four fiction niches I tested, only cozy mystery series was indie-friendly. 4 of 5 top books were self-published. The cost of entry is a 6 to 10 book series. You don't compete in cozy mystery as a one-book debut.

Why fiction needs different methodology

My non-fiction research methodology was straightforward. Sample 10 books from a niche, calculate the median earnings, check star ratings, look at review counts. The opportunity score (median earnings divided by median reviews) was the headline metric.

That methodology fails for fiction in five specific ways.

First, fiction is series-driven. The single-book median in cozy mystery means almost nothing because cozy mystery readers buy in series and Amazon's algorithm rewards prolific authors with deep backlists. The strategic question for fiction is whether you can build a 5 to 10 book series. Asking whether you can write one good book is the wrong frame.

Second, fiction earnings are heavily distorted by Kindle Unlimited. KU pays per page read instead of per copy sold. The standard BSR-to-sales formula was calibrated against traditional sales. For KU-heavy genres like cozy mystery, romance, and thriller, the displayed earnings number understates the actual income an indie author earns.

Third, the review count threshold is different. A non-fiction book with 5,000 reviews is well-established. A fiction book with 5,000 reviews is mid-tier in many genres. Fiction readers leave fewer reviews per copy sold than non-fiction readers, so the same review count represents much higher sales volume in fiction.

Fourth, the rating-gap signal collapses. In non-fiction, a niche where the top books average 4.0 stars (against a more typical 4.5 to 4.7) signals reader dissatisfaction and an opening for a better book. In fiction, almost every successful sub-genre sits at 4.5 to 4.7 stars. The gap signal you'd be looking for barely exists.

Fifth, the publisher mix is a stronger signal than any of the metric scores. A fiction niche dominated by Big 5 trad imprints is closed to indies, full stop, regardless of what the earnings look like. A niche that's mostly indie is open to indies, full stop, regardless of how high the review counts are. None of this shows up in the earnings or opportunity score directly.

I rebuilt the deep-dive tool around all five of these adjustments before running the four niche tests below.

Lesson 1: Generic fiction terms return zero books

I started with what seemed like an obvious search: "small town novel."

Amazon returned zero non-sponsored books.

Not "few results." Not "thin sample." Zero. Amazon collapsed the broad fiction phrase into either sponsored ad placements or non-book product categories.

This pattern hit me twice during testing. Generic fiction descriptors do not function as Amazon search terms the way specific non-fiction problem statements do. "Anti-inflammatory cookbook for women over 40" returned 50+ books in the previous article. "Small town novel" returned zero.

The fix is sub-genre specificity. "Small town novel" is too vague. "Small town romance" works. "Cozy mystery series" works. "Christian literary fiction" works. "Regency romance" works. The pattern is genre plus narrower descriptor, in the language Amazon's category browse pages use, not the descriptive phrases readers might use to talk about a book.

If you're researching fiction, start with the actual category names Amazon uses on its genre pages. Skip the descriptive phrases.

Lesson 2: Earnings math broke for fiction (and the fix)

The first version of my tool computed earnings by averaging the BSRs of the sampled books, then running that average BSR through the BSR-to-sales formula. For non-fiction this gave reasonable numbers. For fiction it produced nonsense.

The first time I noticed this was on a Christian literary fiction run. The displayed result was median earnings of $1,137, average earnings of $28.

The median was bigger than the average. That's mathematically backwards in normal distributions. The bug was structural.

In a niche with 6 books where 3 are bestsellers (BSR around 5,000 to 30,000) and 3 are dead listings (BSR above 200,000), the average BSR lands around 100,000. Plug that into the BSR-to-sales formula and you get about 6 sales a month, or $28 a month at typical fiction prices. Meanwhile the median BSR sits in the bestseller cluster, generating something like $1,137 a month.

The displayed numbers were technically what I asked for. They were not useful information.

The fix is per-book earnings. Calculate each book's monthly earnings individually, then take the median and average of those values. The current tool does this. It matches the methodology used in the original 25-niches article and how most KDP research tools handle the math.

This isn't a fiction-specific lesson. It applies to any niche with high BSR variance. But fiction niches almost always have high BSR variance because the top books in fiction earn dramatically more than the long tail. The bug was much more visible in fiction than in cookbook or workbook niches.

Lesson 3: Publisher matters more than reviews in fiction

My biggest methodological correction came from running Christian literary fiction.

The data: median earnings $2,615 per book per month, median reviews 34,439, top book BSR #141, opportunity score 0.08. Top books included Theo of Golden, The Prayer Box, When Crickets Cry, and The God of the Woods.

A standard non-fiction read of those numbers would say "saturated, skip" because the opportunity score of 0.08 is one of the lowest possible and the 34,439 median reviews suggests a wall of established competition.

That read misses the actual reason this niche is dead for indie authors. The Prayer Box is published by Tyndale Fiction. When Crickets Cry is published by Thomas Nelson, a HarperCollins imprint. The God of the Woods is published by Riverhead Books, a Penguin Random House imprint. These are traditional publishers with bookstore distribution, marketing teams, and backlist catalog placement that no indie author can match. The high review counts are a downstream effect of trad publishing. The structural lock is the underlying cause.

A new indie author writing the best Christian literary novel of 2026 cannot rank in the top 10 of this niche. The writing quality has nothing to do with it. The distribution channel does.

The headline metric for fiction niche research should be the publisher mix in the top 5 or top 10 results. If 3 or more are Big 5 trad imprints, the niche is closed to indies. If 3 or more are independently published, the niche is open. The earnings and review numbers are secondary.

I rebuilt the tool to extract publisher data from each book's PDP, classify it as Big 5 trad, mid-size trad, indie, likely-trad, or unknown, and surface the mix at the top of every analysis result. Each top book now shows a colored badge so you can see at a glance whether the niche is competable.

Lesson 4: Kindle indie books have no Publisher field at all

Building the publisher detection turned up something I didn't expect. For Kindle ebooks where the author is the publisher, Amazon doesn't show a Publisher field. The metadata is just absent.

I confirmed this by pulling the raw HTML for the top cozy mystery result, "Tiffany Black Cozy Mystery Compilation Books 1 to 12." The HTML had eight book detail tokens for things like file size, word wise, page flip, and X-Ray. No publisher token at all. The string "Independently published" wasn't in the HTML anywhere.

The reason is Amazon's display logic. When the publisher is a real publishing house, Amazon shows it. When the author is the publisher, Amazon omits the field rather than displaying "Publisher: Yourself." Kindle indie books just have no publisher data, by design.

This is actually useful. The absence of a publisher field on a Kindle book with moderate reviews is a strong signal that the book is indie. I added this as a heuristic to the classifier. If the publisher is missing AND the format is Kindle AND review count is below 5,000, classify as indie. The 5,000 ceiling avoids mislabeling long-running successful indie series.

For paperback indies the situation is different. Many indie KDP paperbacks do show "Independently published" as the publisher, because Amazon needs to show something on the print listing. The classifier catches those through the "Independently published" string match.

The composite logic catches both cases. In the cozy mystery test that followed, four of five top books were correctly tagged indie, two via the "Independently published" string and two via the Kindle-without-publisher heuristic.

Lesson 5: The 4.0-star opening doesn't show up in fiction

In the non-fiction research, the strongest opportunity signal I found was a 4.0-star rating average against a backdrop of 4.5 to 4.7 in the rest of the category. That meant readers were buying the available books, leaving disappointed reviews, and waiting for someone to do it better.

In fiction, that gap doesn't show up.

Christian literary fiction averaged 4.7 stars across the top 5. Cozy mystery series averaged 4.5. Book club fiction averaged 4.3. Even the 4.3 wasn't really an opening. The book club niche was dominated by Reese's Book Club picks (Eleanor Oliphant, Something in the Water, Broken Country) and the 4.3 reflected the polarizing nature of those titles, not a quality gap a new author could exploit.

Fiction readers behave differently from non-fiction readers. They give 5 stars to authors they already like and 1 star to books that broke their expectations of a series or trope. The middle ground is thin. Looking for 4.0-star niches in fiction means looking for niches where readers are systematically dissatisfied, and that pattern is rare because readers either love a sub-genre or don't read it at all.

Star rating still matters in fiction analysis. Use it to confirm a verdict rather than form one. If a niche has 4.5+ ratings and is mostly indie, that's a healthy market. If it has 4.5+ ratings and is mostly trad, it's a closed market that readers happen to like. The publisher mix is doing more work than the rating.

The 4 niches I tested

Christian literary fiction: skip

Sampled 8 of top 10 books. Median earnings $1,453 per month per book, median reviews 5,165, average rating 4.7, opportunity score 0.28. Top book BSR #990. Average price $10.01.

Three of the top five were traditionally published. The Prayer Box (Tyndale Fiction). When Crickets Cry (Thomas Nelson, a HarperCollins imprint). The God of the Woods (Riverhead Books, a Penguin Random House imprint). Two were unknown.

This niche is dominated by traditional publishing. Even a strong 5-book series from a new indie author will not crack the top 20, let alone the top 10 where the meaningful earnings concentrate. Skip unless you already have a trad deal or an existing audience.

Book club fiction: skip

Sampled 5 of top 10. Median earnings $32,572 per month per book. Median reviews 19,906. Average rating 4.3.

The earnings number looks insane. Then look at the top books: Eleanor Oliphant Is Completely Fine, Something in the Water, Broken Country, The Calamity Club, The Book Club for Troublesome Women. Every single one is a Reese's Book Club pick or otherwise traditionally published bestseller.

The $32K median is real income. It is not income an indie author can reach in this niche. Writing quality has nothing to do with the barrier. The barrier is access to Reese Witherspoon's marketing pipeline. Skip.

Small town novel: not a real Amazon search

Sampled 0 of 10. Amazon returned zero non-sponsored books for this keyword.

The keyword is too generic. Amazon collapsed it into ads or non-book results. This niche cannot be analyzed without a more specific term. If you're targeting the small-town fiction reader, search "small town romance" or "small town mystery series" instead.

Cozy mystery series: indie-friendly with caveats

Sampled 9 of top 10. Median earnings $414 per month per book, average $9,395 (the gap means there are huge outliers like the Tiffany Black 12-book compilation). Median reviews 2,291. Average rating 4.5. Opportunity score 0.18. Top book BSR #45.

Four of five top books were independently published. Two had "Independently published" in the publisher field. Two were Kindle ebooks with no publisher field at all (the indie heuristic caught them). Only one book in the top 5 had an unknown publisher (Marsh Mystics, J&R Publishing).

This niche is indie-friendly. The competition is other indies, not trad publishers. The cost of entry is volume. The Tiffany Black compilation is "Books 1 to 12." Pelican Cove is a "Series Box Set." The cozy mystery algorithm rewards prolific authors with deep series and Kindle Unlimited backlists. Writing one cozy mystery and walking away will get you nowhere. Committing to a 6 to 10 book series over 18 to 36 months is the realistic strategy.

This is the only one of the four niches with a genuine path to indie income. The path requires sustained output instead of a single hit.

How to do this yourself

If you're researching a fiction niche before committing to a series, this is the workflow that actually works.

Start with sub-genre specificity. Use the names Amazon uses on its category browse pages. "Regency romance" works. "Period romance set in England" doesn't.

Sample at least 10 books per niche. The top 3 results in fiction are even more survivor-biased than in non-fiction because fiction's bestseller distribution is more extreme. Books 6 through 10 reveal the long tail honestly.

Compute median earnings, not average. Averages get dragged up by the few mega-hits in every fiction niche. The median tells you what your book is realistically going to do.

Check the publisher mix in the top 5. This is the headline metric for fiction. If 3 or more are Big 5 trad imprints (Penguin Random House, HarperCollins, Simon & Schuster, Hachette, Macmillan) or major digital-first trad imprints (Bookouture, Lake Union, Thomas & Mercer, Aria Fiction), skip the niche. If 3 or more are independently published or Amazon-Publishing-imprint titles, the niche is open.

Verify the niche actually exists. Search Amazon directly with your keyword. If you get zero non-sponsored books, the keyword is too broad. If you get a top 5 dominated by Reese's Book Club picks or NYT bestsellers, the niche is uncompetable regardless of what the earnings number says.

Plan for series. A fiction niche analysis is asking whether you can commit to writing 5 to 10 books here, instead of whether this is a good single-book niche. If you can't commit to that volume, find a different lane.

You can do this manually with Amazon search and a spreadsheet. Plan for 8 to 12 hours per round of 10 niches.

Or you can use a tool that runs the whole analysis in 30 seconds.

The tool I built (fiction-aware now)

BookRankPro does the fiction analysis automatically. The Deep Niche Analysis page has a fiction-vs-non-fiction toggle. Select fiction and the tool applies fiction-specific reasoning to the verdict, factors in the publisher mix as the primary signal, and uses adjusted thresholds for review counts and competition scores.

Each top book gets a colored badge: red for Big 5 trad, orange for mid-size trad, amber for likely-trad, green for indie, gray for unknown. The publisher mix card at the top of every result tells you in one sentence whether the niche is indie-friendly, mixed, or trad-locked.

It's a $97 one-time payment, no monthly subscription, 30-day money-back guarantee. The same price as before, with the fiction features now included.

If you want to skip the rabbit hole I went down (the broken earnings math, the missing publisher metadata, the small-town-novel zero-results dead end), that's what BookRankPro is for.

Key takeaways for fiction KDP authors in 2026

Don't search with generic terms. "Small town novel" returns zero books. "Small town romance" returns 50. Specificity matters.
Median earnings beats average earnings by a wider margin in fiction than in non-fiction. Use the median.
Publisher mix is your headline signal. If the top 5 has 3 or more Big 5 trad imprints, the niche is closed to indies regardless of what the earnings or reviews look like.
Kindle indie books have no publisher field. The absence is itself a signal once you know to look for it.
Fiction is series-driven. Plan for 5 to 10 books. If you can't commit to that volume, the genre isn't for you.
Cozy mystery, romance, and thriller are the indie-friendly fiction lanes. Literary fiction, book club fiction, and most prestige sub-genres are trad-locked.

FAQ

What's the most indie-friendly fiction niche on KDP right now?

Cozy mystery series, based on my testing. 4 of 5 top books are independently published. Median earnings around $414 per month per book, but the average is much higher because of box-set compilations from prolific authors. The barrier is volume. You need 6 to 10 books to compete with established indie authors. Romance and thriller share similar dynamics but I haven't tested them with the new methodology yet.

Why does literary fiction not work for indies?

Trad publishing has a structural lock on literary fiction through bookstore distribution, NYT review coverage, book club partnerships (Reese's, Oprah, Read with Jenna), and prestige award circuits. None of those channels are accessible to indie KDP authors. The top books in literary fiction are dominated by Big 5 imprints because that's how literary fiction gets discovered, not because the writing is structurally better.

Are AI-generated fiction books profitable on KDP?

Untested in this round. My non-fiction research found that profession-specific AI prompt collections didn't actually exist as a real KDP category. AI-assisted fiction is a different question. Many indie authors use AI to draft or polish but write the actual book themselves. That's distinct from fully AI-generated fiction, which Amazon has been quietly cracking down on with content quality flags and disclosure requirements.

What's the typical earnings range for an indie cozy mystery book?

Wildly variable, like all fiction. The bottom quartile makes under $50 a month. The median makes around $400 to $1,200. The top books with established 10-plus-book series make $5,000 to $15,000 a month, mostly through Kindle Unlimited reads. Aim to match the median in your first 3 books and the top quartile by book 6 if you stay consistent.

Do I need a series to compete in fiction?

For most fiction sub-genres, yes. The Amazon algorithm rewards series velocity (how fast you publish) and series depth (how many books in the series). Standalones can win in some literary niches but lose in cozy mystery, romance, thriller, and most genre fiction. If you can't commit to writing at least 3 books in the same world or with the same characters, fiction may not be the right format for you.

How is BookRankPro different from Publisher Rocket for fiction research?

Publisher Rocket gives you keyword data and competition estimates. BookRankPro gives you live BSR, publisher classification per book in the top 5, the publisher mix as the primary fiction signal, and a Claude-powered verdict that factors all of it together. The publisher detection is the bigger differentiator. No other tool I've found classifies whether a top book is Bookouture vs Tor vs Independently Published. That's the metric that actually predicts whether an indie can compete.

This research was conducted on May 2 and May 3, 2026, using live Amazon data via BookRankPro's Scrape.do integration. All numbers reflect Amazon results at the time of testing. Niche data shifts week to week. For current numbers, run your own search.

Author

Obi Anthony is a first-time KDP author and the founder of BookRankPro, a live-data niche research tool for indie KDP publishers. He lives in Atlanta, Georgia, where he is currently working on his debut children's handwriting workbook.