KDP niche research: how to find a profitable niche by the numbers

Why the niche matters more than the writing

A book on Amazon is a product, and it follows product rules. A beautifully written book in a crowded niche will sell worse than an average book where there is demand and few competitors. The niche is step one; everything after it, from the book itself to ads, is in our guide on how to sell books on Amazon KDP.

The gap isn't 10 or 20 percent. It's several times. The same work — text, cover, ads, the same budget — gives steady sales in a live niche and almost nothing in a saturated one. And you don't find out right away. You find out three or four months later, when the ad money is already spent.

A cover can't fix a bad niche, and neither can ads. Ads just show a weak product to more people faster, and burn the budget faster.

So you check the niche before you write a single line.

The 5-minute niche check

Before any deep research, you can filter out most ideas in five minutes.

1. Search demand. Go to Amazon.com and start typing your topic into the search bar. If Amazon suggests completions, people are searching for it. No suggestions is a warning sign.

2. The top 10 books for the keyword. Open the first ten results and check two things: how many reviews they have and their Best Sellers Rank (BSR). If most of them are under BSR 40,000 with 50+ reviews, the niche is active.

3. Covers. Do the covers in the results look current? Are competitors updating their books? If the top sellers look dusty and dated, the niche has probably stalled.

4. Publication dates. If everything on top is from 2020–2021 and nothing new has broken through since, the market is frozen. Fresh books with good ratings in the top mean the niche is alive.

5. Compare against this table.

Do this before you publish and you'll skip most of the disappointments.

Step 1. Look at what already sells

Don't start with "what should I write about?" Start with "what are people already buying here?"

Open the category on Amazon and write down 10–15 books similar to the one you want to make. From here on, you research these books, not an abstract topic.

Two rules, without which the research is meaningless:

Your competitors are independent publishers only. Ignore books from traditional publishers entirely: different budgets, different contracts, different promotion. You are competing with other independent publishers like you.

Look at organic results, not ads. A book in first place with a Sponsored label is there because someone paid for it. That says nothing about demand.

Step 2. What to check in each competitor

BSR: the key number

BSR (Best Sellers Rank) is the most important number on the page. The lower it is, the better the book sells. A book at BSR 5,000 sells far more than a book at BSR 200,000.

The screenshot shows three books from one search: BSR #85,614, #3,522 and #13,695. The first barely sells; the second and third are alive. Product details shows what matters most: Independently published — a self-publisher like you, not a publishing house. Plus the page count (78 pages) and the category ranks.

Our three benchmarks:

• BSR under 40,000 — the niche is alive, people buy books in it.
• BSR under 20,000 — our working threshold for entering with a new book.
• BSR over 100,000 with 0–20 reviews — the niche is most likely dead.

The difference between the first two matters. A niche at BSR 30,000 is alive, but harder for a newcomer: established players with reviews are already there. A niche with books under 20,000 gives you more room to climb.

Look at the whole top, not one book. If it's all publishing houses, you're too early for that niche. If there are independent authors under BSR 40,000, there's room for you.

What to check besides BSR

Number and freshness of reviews. A book with 1,000 reviews over five years and a book with 100 reviews in six months are different stories. The second means the niche is alive and new players are getting in.

Price and page count. Compare competitors' prices and lengths. If everyone in the niche sells 200 pages for $15, your 75-page book at $24.99 loses at first glance — even if it's better.

Check right away whether a profitable book is even possible at the niche's prices: KDP pays 60% of list price minus printing costs on paperbacks priced $9.99 and up (KDP Help), so printing eats into every copy. The full math is in KDP royalties explained.

Negative reviews. The most useful thing on a competitor's page. Readers tell you exactly what they missed: not enough examples, tiny print, filler, no index. That's a ready-made brief for your book.

Step 3. Tools

AuthorLens is an analytics dashboard made only for Amazon KDP.
We use AuthorLens daily and have a business relationship with it.

For niche research, the relevant part is its competitor, niche and category screens — Books, Compare, Discover and Opportunities — where you can look at books and categories side by side instead of opening every listing. After the book is out, the same dashboard tracks daily rankings for up to 500 keywords and shows Keyword ideas: words buyers used to find you that aren't in your listing yet. It covers Amazon only, and that's its weaker side if you also sell elsewhere. More in our AuthorLens review.

How to read the Opportunities screen. It lists Amazon categories and shows, for each one, whether a new book can realistically reach the top 20:

• Sales at #20: an estimate of how many copies a month the book in 20th place sells. This is the demand you are competing for.
• Reviews to get in: how many reviews that book has. It is one book, so it can be a fluke in either direction.
• Top-20 reviews: how many reviews a typical top-20 book has. Read it next to the previous number.
• Chart movement: the share of books that changed place. A high number means new books do get in.
• Average price: what readers in the category are used to paying.

Look for a category where sales at #20 are high and both review numbers are low. In the screenshot, two romance categories sell about the same at #20 (5,572 and 4,474 copies a month), but the book at #20 has 32,591 reviews in Small Town Romance and 46 in Mafia Romance; a typical top-20 book has 2,076 against 966. The sales figures are estimates, so compare categories with each other rather than treating a number as a promise.

Free tools that help at the start:

DS Amazon Quick View — a Chrome extension that shows BSR right in the search results, so you don't have to open every book. Saves hours.

Google Trends — shows general interest in a topic over time. Remember it's Google search data, not Amazon, and the two can differ. It's good at catching seasonality, but it doesn't show real demand in the book market.

When a niche looks promising and you want to check it seriously, paid tools help:

Helium 10 — keyword, niche and competitor research. Promo codes from Vova Even: VOVA10 — 10% off for life, VOVA6M20 — 20% off the first six months. Which parts of Helium 10 actually work for books is in our Helium 10 for KDP guide.

Publisher Rocket — a one-time payment ($199 on its site as of September 2026); analyzes sales, rankings and keywords specifically for books. Details in our Publisher Rocket review. For a side-by-side of everything, see best KDP tools.

Step 4. Signs of a dead niche

A niche can look attractive and still be closed to new players. Warning signs:

• The same books sit on top for years. New ones don't rise — the market is locked by old players with piles of reviews.
• New books have almost no reviews. People come in, publish, and get no sales.
• Prices are sliding. Competitors are undercutting each other — a sign sales are falling and they're trying to rescue them with price.
• Ads get more expensive while sales don't grow. The costliest signal, because you notice it after you've spent.

If several of these match, look for an adjacent topic instead of hitting a wall.

Step 5. The most expensive mistakes

Choosing a niche because you like it. Your interest matters: in a topic you understand, it's easier to brief a writer and spot shallow text. But interest doesn't replace demand. Numbers first, taste second.

Choosing by one metric. Low competition alone means nothing — maybe nobody buys there. Look at demand and competition together.

Going where it's easy to get in. Coloring books, word searches and planners attract people because there's nothing to write. That's exactly why they're crowded, and the margin is thin enough that ads don't pay off: are coloring books profitable on KDP.

Entering a topic where the competitors are publishing houses. You lose not on quality but on budget and accumulated reviews.

Building a conveyor of ten books. "If one doesn't sell, the next one will" usually gives you ten weak books instead of one strong one. Our account with just three books had a 22% ACOS (ad spend as a share of ad-attributed sales) — the lowest of all our accounts — because the niche was hit precisely.

Not running the numbers before writing. The most expensive mistake of all. A 200-page full-color book at $15 can't carry ads: in the US, standard color on a regular trim costs $1.00 plus $0.0255 per page to print (KDP Help), which is $6.10. Sixty percent of $15 is $9.00, so you keep $2.90 a copy — less than it usually costs to win one buyer with ads. Our KDP royalty calculator shows this in two minutes, before you pay for a single word.

Seasonality: a niche that lives three months

Check separately whether your topic is seasonal. BSR on the day you look won't tell you.

Gift books, holiday themes, yearly planners, Christmas coloring books — they spike in Q4 and go quiet the rest of the year. You'll see great numbers in November and think you've found a gold mine. In February sales drop several times over and you'll have to stop the ads.

The check is simple: look up the topic in Google Trends over two or three years. Clear peaks in the same months every year mean a seasonal niche.

That's not a reason to walk away. You can take a seasonal niche on purpose: launch in spring, collect reviews by fall, harvest in Q4. Just budget for nine months of the book roughly breaking even.

The problem is when seasonality comes as a surprise.

Trend vs. evergreen

Seasonality repeats every year. A fad doesn't repeat at all, and that's more dangerous.

An evergreen topic is one people will still search for in three years: nutrition, habits, exam prep, houseplant care. A fad lives a few months: a new trend, a hit TV show, a diet everyone is talking about right now.

The catch is that a fad looks better than an evergreen topic when you check it. Big demand, few competitors, great BSR on top — it feels like the niche of your dreams. But while you write the book, make the cover and start ads, two or three months go by, and the wave is gone.

Our own example. We pulled one book from sale after six months: it made about $200 in profit. Partly it was price — at $20.99–24.99 nobody bought it, and at $18.99 ads ate all the profit. But the root cause was the topic: it turned out to be a fad, not evergreen. The hype faded fast, and we apparently missed the peak.

How to tell them apart. Google Trends again, but look at the shape of the line. Seasonal — even peaks in the same months year after year. Evergreen — a flat line with no sharp moves. Fad — one tall peak and a drop that never comes back.

If the line shot up in the last few months and never did before, it's a fad. You can go in, but only with a book that's ready now and the understanding that the window will close.

Look beyond cookbooks

One thing we've come to over the past year.

Cooking is the most popular way into the book business, and that's exactly why it's the most crowded. Everyone checks the same niches, fights for the same keywords and bids each other up.

Outside cooking, Amazon has dozens of categories with noticeably less competition. Here is June 2026 on one such book:

Look at the ACOS: 41% is far from perfect advertising; there's still plenty to optimize. And the book still made over $1,600 profit in a month. In an overheated niche, the same result would cost a lot more.

So when you check topics, don't drop directions outside food: psychology, self-help, narrow practical topics. That's often where there's room.

Our own failure, in numbers

So it doesn't sound like niche research is a formality, here is one of our own mistakes.

The Galactic Adventures of a Little Girl — motivational stories for kids. We made it in four language editions at once: Ukrainian, English, Italian and German. The logic was: more languages, more markets, more sales.

35 copies and about $175 in royalties, lifetime. Four covers, four uploads, four listings — and a result that didn't pay for the work on even one edition.

What went wrong: we picked a topic we liked instead of checking demand. Translations don't save that. If nobody searches for a book, nobody searches for it in any language.

The five-minute checklist above would have shown it before we started.

When the niche runs out under you

The first case was a niche that never existed. This one is harder: the niche was real, it worked and made money, and then it started to fade.

We wrote two books on the same topic on different accounts. Both made $2,000+ in profit a month — exactly the result this whole business is for. The logic was obvious: the topic works, let's make a third.

Here's the third book:

$15,728 in royalties over a year — about $1,300 a month. That's a working book, not a flop. But look at what those sales cost:

What was left as profit

Royalties of $15,728.22 minus $9,991.04 in ads leaves $5,737 for the year, about $480 a month. Ads take 63.5% of royalties, and an ACOS of 50.17% is twice what we see on our healthy accounts.

This is what a fading niche looks like in numbers: the book sells, there are almost 1.8 million impressions, orders keep coming — but profit is thin because every buyer now costs more. Why ACOS alone misleads here is covered in Amazon Ads for authors: profit, not ACOS.

Demand in the niche has started to fall. Not collapsed — started to fall. What used to sell on its own now has to be bought with ads — it shows up as a falling organic share: organic vs ad sales on KDP. Every month means new marketing moves, keyword reviews, listing changes and new reasons why someone should pick this particular book.

We aren't giving up: the goal is to take it from $480 to $2,000 profit a month, and it's a daily job. The main lever isn't sales, it's ACOS: bringing it from 50% down to 30% at the same sales would save about $4,000 a year without touching the book. But the lesson matters more than the book.

What follows from it

A niche isn't forever. A niche that worked last year and works now may not work a year from now. Check it regularly, not once before launch: are your competitors' BSRs slipping, are new players coming in, do new books get fewer reviews?

And separately: your own success in a niche is not a check of the niche. Two books that made $2,000+ didn't guarantee the third would. Every new book gets checked by the numbers again, even in a topic you already know.

If the niche doesn't pass

The usual reaction is to get upset and start over. In fact, a check that said "don't go in" just saved you the $3,700–4,700 a launch costs, and several months.

What to do next:

Look for an adjacent topic. A niche is rarely dead as a whole. Usually the center is saturated, while narrower searches on the side still have live demand. Instead of "cooking," a specific diet; instead of "planner," a planner for a specific job.

Narrow the audience. The same topic for a specific group of readers is a different niche with different competitors.

Look for what existing books are missing. Competitors' negative reviews often reveal a niche inside the niche: the topic exists, but everyone does it badly.

Then run the new idea through the same checklist. Five minutes of checking is cheaper than six months on a book that doesn't sell.

What to do next

A chosen niche isn't a topic name. It's a concrete answer to three questions: who the book is for, how it differs from what's already in the results, and how much it can earn at current prices and printing costs. That answer becomes the brief for the book: how to create a book for KDP.

If you can answer all three with numbers, move on to publishing. If even one is missing, spend another hour checking rather than months on the book.