Book a free discovery callBook a free discovery callThe Fall 2026 Reading Report
Every fall the author group chats spiral: sales dip, KU payouts wobble, and suddenly everyone wants to rewrite their backlist or quit. Before you burn your platform down, read the numbers with me: the plays hiding in the data, the mechanical reasons a solid book can look like a flop, and my read on every chart.
Step away from the KDP dashboard. I brought the spreadsheets.
What readers are searching, borrowing and anticipating this fall, with my read beside every chart.
Not sure where to start?
“Kindle unlimited” search flatlined while “audiobooks” climbed to meet it and reading apps hit a five-year high. Audio demand merits a costed test where your readers actually listen.
Romance holds the top of book-specific search 365 days a year, with fantasy climbing steadily behind it. Horror gets its clean autumn surge, then drops off a cliff in November.
Fourth Wing and the ACOTAR series hold a massive baseline all year. The fall slate mixes deep-series sequels with debuts, which leaves real room for an accessible book one.
Carousels and Reels lead on saves in this cross-industry sample, so test formats against your own audience. Write captions like search queries, and keep a real human face on your content.
Direct sales and an email list are where the real money gap shows up. Build the list before you need it, so you never have to beg an algorithm for reach.
Every big platform is pushing AI at the casual reader, not the one in your author group. Be the book a machine can describe in plain words, and the human it can point to.
When sales go quiet, check where readers drop off: discovery, the sales page, or the book itself. Fix the step your own numbers point to before buying more reach.
Pattern-spotter. Professional over-thinker.
The human behind the spreadsheets
Confession: I am obsessed with data.
I run Literary Inspired. Early on I realized I was unreasonably good at spotting patterns in data and turning them into systems. Turns out that was undiagnosed ADHD with a dash of OCD for good measure. These days I wield my neurospicy talents like a seventeen-year-old FMC about to save the kingdom, from the safety of my desk, behind a screen, because, germs, obviously.
That is gold for a strategist reading consumer habits. It means I will not make a recommendation I can’t back with a hard number. So I brought the data, and the diagnostic questions to ask before you look at a quiet dashboard and decide the answer is another book, another platform, or another job for yourself.
The subscription economy that pays your page reads, and where reading demand is moving next.
This is the money chapter. It covers how Kindle Unlimited actually pays, where reading demand is drifting next (audio, libraries, reading apps), and what both mean for a book you are deciding whether to enrol.
The data here is search interest and the public fund tracker, so it shows demand and platform economics, not your personal page reads. Read it to make two calls: KU or wide, and whether audio earns a place in your plan this year.
The total pot Amazon pays out each month, against what it actually pays you per page read. Watch them move apart.
Toggle fund, payout, or both together.
The fund keeps growing. What it pays you per page has not.
Kindle Unlimited interest has flattened while audiobook search keeps climbing.
Where reader attention is drifting across Kindle, audiobooks and reading apps over time.
Switch between the 5-year and 12-month view.
Spotify is the biggest term in this audio comparison, and that includes music.
Readers look for books in more places than Amazon.
In any given week, books lose the fight for attention.
A slow launch gets blamed on marketing by default. Check read-through first. If book one turns strangers into readers and book three loses them, the likely culprit is fit between the books (positioning, series order, even price), so check those before you spend on traffic.
Data extracted September 18, 2026
It is tempting to read falling KU searches as readers leaving. The fund tells a different story: it was still near a high, around $67.6M a month in July 2026 (and higher since), even as branded search fell.
The most likely explanation is brand familiarity. People may simply know what Kindle Unlimited is now, so they search the words less. The chart shows the drop, not the reason, so hold this as the strongest read rather than proof. The “readers want to own, not subscribe” theory is the weakest fit, because the fastest-growing format is audio, which is almost all subscription.
The quieter squeeze is per-book.
So where did my page reads go
A 2025 rule change a lot of authors missed: Amazon updated the KDP Select terms so your enrolled ebook can now reach public libraries during the exclusivity period, without leaving Kindle Unlimited.
Why it matters: Library borrows put your book in front of readers who would never have found it on Amazon, and they do not cost you your KU page reads. I have written the full how-to here: KDP Select Now Allows Library Distribution for KU Books.
The genres readers search all year, and the calendar those searches follow.
This chapter answers two questions: which genres readers are searching for, and when across the year they search for them.
How interest in the main fiction genres compares, and how that ranking shifts over a longer window.
Switch between the last 12 months and the 5-year shape.
The reading calendar repeats, almost to the month.
Romance peaks for summer reading and again at the holidays. Horror concentrates into one sharp October block. Fantasy and thriller both climb hard into December, while mystery and young adult stay steadier all year.
The old seasonal assumptions are worth testing. Here is what the evidence actually supports, and where it does not.
January: lots of searching, little buying.
What the data says: The buying is front-loaded into December. Over 40% of the year’s print and ebook sales happen between December 1 and January 15, so January reads quiet at the till while readers are busy reading: finishing gifted books, loading new Kindles, and signing up for a reading challenge in droves (Goodreads logged 316,709 sign-ups by January 1, 2023). For you: January is a “what do I read next” month. Be the easy answer, with a visible series starter and a clear next-book link in the back.
May: a softer month for sales.
What the data says: In one recent May, reporting-publisher sales fell 7.5% year on year. One month is not a trend, so treat it as a flag to check your own May numbers, not a rule. The usual “everyone is busy with graduations” story is not in the data either. May reads more like a shoulder month, the lull between spring and the summer beach-read surge, so readers have not left, the calendar has just moved them. For you: If your genre rides the summer wave (romance, thrillers, anything beachy), May is better spent on setup and backlist than a launch; if you write to a different calendar, May may be fine for you.
The holidays: short, fast, finishable.
What the data says: Busy readers do reach for shorter, faster books, and December piles it on: no time to read, and a Goodreads goal to close before the year ticks over. Goodreads runs dedicated year-end readathons and “short books to hit your goal” lists for exactly this reader. Print also spikes now, because a physical book gives better. For you: A novella, a quick holiday read or a boxed set earns its place in December. Call it the one book they can actually finish before New Year.
Does one format own a season?
What the data says: Yes, and predictably. Audiobooks peak in summer with travel and commuting (73% of listeners listen while travelling, and audio climbs to its high around July). Ebooks get a January bump from gifted e-readers. Print owns the November and December gift window. For you: Match the format you push to the month. An audio release has its best shot heading into summer, not mid-winter.
These are directional patterns, drawn from sales reporting, Google Trends, Goodreads challenge data and audiobook listening studies. The reading-calendar heatmap compares six genres (romance, fantasy, thriller, mystery, horror and young adult) from US Google Trends, 2022 to 2025. Each genre’s months are scored against its own yearly average, and each row is then shaded across its own range, so darker cells mark a genre’s stronger months rather than comparing one genre to another. Other Trends charts in this report use the geography stated on each. Exact figures and sources are on the Sources tab.
The books readers are counting down to between now and February, and what a release this size does to everyone else’s reach.
This is the company you keep this season: the books readers are already counting down to between October and February. Dates are the confirmed US edition, and every cover links to its book page. The early-October titles are already out. The rest land between now and early 2027. Use the filters to find who you are launching beside.
Confirmed releases
Two ACOTAR books land in this very window (October 27 and January 12). When a release is this big, authors feel their own reach dip and assume readers vanished. They did not. Here is the real mechanism, and what you can do about it.
The creators move first.
What happens: BookTok and Bookstagram pivot hard to the big release, because a meme or a reel about it is the one most likely to go viral. When A Court of Silver Flames dropped, that wave was enormous. The casual reader who usually discovers books through those creators suddenly has her feed pointed somewhere else.
Nostalgia pulls lapsed readers back.
What happens: Readers who loved ACOTAR in their late teens get curious all over again. For a lot of people (me included) that series was the gateway back into reading at all. A big sequel reactivates the whole fandom, lapsed readers very much included.
A divisive book is a louder book.
What happens: Silver Flames handed the spice a notch and gave readers Nesta: prickly, messy, mid-healing, a bit of a bitch, and nothing like agreeable Feyre. Whether people loved her or hated her, they talked. Everyone wants in on the conversation, the memes and the inside jokes.
Then comes the homework.
What happens: The new arc rewards readers who know the connected universe, which runs to roughly 16 books across three series. Many fans re-read or catch up rather than risk spoilers, so attention pools hard around the launch. A newcomer does not have to read all three series to enjoy the new book, but the gravitational pull on the genre is real. Think Swifties, for books.
So readers did not disappear, their attention moved
Throne of Glass, ACOTAR and Crescent City now overlap. New readers face the whole shelf before the new arc lands, which is why the re-read wave is so strong right now.
Series openers and standalones on this slate start readers clean. That accessibility is their advantage against the giants, and yours.
Long series build loyal fandoms. The catch-up they demand is also what leaves a door open for an accessible book one.
Books are bought by consumers, and the consumer has changed. What survives the reader’s filter, and where they quietly drop off before they buy.
This chapter is about how readers find, judge and buy books now, and what you can change at each step. It opens on AI because that is the ground shifting under everything. This year every major platform (Google, Canva, Gemini, Meta, Apple) said the same thing out loud: AI integration is the priority.
Picture the casual reader, not the one in your reader group: the busy mom reading between school runs, the guy who gets fifteen minutes before bed, the gran scrolling Facebook for photos of the grandkids. She never sees the discourse. She just sees a “tool” added to every app she already uses, and before long, convenient starts to look like frozen meals for every dinner. So yes, the data shows readers are already more likely to ask a chatbot what to read next.
Remember,
Your reader is not debating AI, she is being handed it: So your job splits in two: be the book a machine can describe in plain words, and be the human it can point to. The charts below are how she decides once the machine hands her the shortlist.
of consumers will not buy a product they can afford if its value feels weak. Gen Z leads a barbell pattern: splurge on what matters, trade down elsewhere.
Most readers move through these six steps, some fast, some over weeks, and a few skip straight to buy on a trusted recommendation. Most author marketing still pours everything into the first step.
Why it matters: When sales are soft, the break is usually further down than Notice. Find the step where your readers fall away, and spend your energy there before buying more eyeballs for a book that leaks.
Authors can tell us what they try. Readers tell us what survived the filter.
Readers trust taste-match over mass approval.
What it suggests: They seek reviewers who share their taste and read the 1 to 3-star reviews for dealbreakers. A specific three-star review can sell the book to the right reader; stop treating every non-five-star as a failure.
Tropes qualify readers but no longer differentiate.
What it suggests: “Enemies to lovers, morally grey, one bed” now reads as interchangeable. Position in two layers: first, does this belong in my reading world; then, why this book over the others with the same tags.
Reader requests are getting hyper-specific.
What it suggests: People ask for exact tone, heat, age and “do not want” lists, not “fantasy recs.” Reader avatars should go past demographics into that specificity, which then feeds hooks, comps, ARC screening and ad copy.
AI suspicion changes how readers evaluate risk before opening a book.
What it suggests: They check samples, three-star reviews and the author's human presence. The answer is not shouting “human written,” it is looking unmistakably cared for. Note: readers also misattribute ordinary quality problems to AI.
The ARC community is having a loud, useful argument right now.
Raw signup count is almost the wrong ARC metric. Every stage leaks. A handful of well-matched readers usually outwork a big indiscriminate list.
On the KDP wobble: KDP reporting has been erratic in September 2026: spikes then zeros, widely read as delayed, non-real-time reporting rather than lost reads, with a genuine ACOTAR re-read wave possibly pulling reads toward the backlist. Watch weekly and monthly, not the daily number.
“Too much dark banter, not enough sweet romance. The heroine never softens and the jokes get mean.”
Add to cartThe dark-banter girlies, adding to cart.
Treat the algorithm as the last thing to blame. Run the split, then do the five-minute check: open your buy link on mobile, confirm the format and region show correctly, read your own sample, and make sure the next-book link works.
Raw sign-ups flatter the ego more than they move the needle. Forty genre-matched readers who finish and review will usually do more than four hundred who signed up for a free book and never opened it. Recruit for fit, keep the ask simple, and follow up like a human.
What the 2026 trend pieces broadly agree on, and what actually moves the needle on social for an author with limited hours.
This chapter is the wide-angle view: what the 2026 trend pieces broadly agree on, and what actually moves the needle on social for an author with limited hours.
The social figures are platform studies drawn from mixed industries rather than book-specific panels, so read them as direction, and weigh each against your own numbers.
The indie ground is still growing, and it pays best to authors who own their readers.
The global audiobook market tops $10 billion and grows double digits. Getting found inside AI answers is predicted to overtake search, and about 40% of US adults have stopped reading for pleasure.
Where readers spend attention, and what actually performs there now.
of social pros now use AI to write captions, yet a growing share think AI content performs worse. Engagement rate matters less than it used to. Readers reward audience alignment and real storytelling over raw reach.
Authors agonize over the blurb, then leave backend categories and keywords on autopilot, and the book sinks to page forty-seven of the store’s search results, a shelf no reader scrolls to. Metadata decides whether your book enters the room. The blurb only closes the sale once someone finds it.
Carousels earn saves and Reels bring reach. A single static image still works for a quick announcement, it just travels less far.
The parts that touch your book, your covers, your ARC team and your audiobook, with the real numbers and the real caveats.
This is the chapter everyone is anxious about. Read it to tell the changes worth acting on from the noise worth ignoring. The numbers are sourced, and where a figure is soft or contested, it says so.
About 4 million books were published in the US in 2025, up 32.5% in a year. Self-published titles jumped nearly 39% to more than 3.5 million, outnumbering traditional roughly five to one.
AI did not start self-publishing, but it poured fuel on the volume. The takeaway is not “write more to keep up.” It is that discoverability and trust now do the work sheer output used to.
You may have heard “two uploads a week.” The accurate version: Amazon limits you to two new titles per format, per week (so two ebooks, two paperbacks, two hardcovers).
Publishing one book with its paperback and hardback stays well under the wall. The cap targets mass-publishers flooding the store, and it tightened from three titles a day in 2023 to this, in direct response to AI-generated flooding.
In one widely-shared reader poll, about 80% said they would not buy a book with an AI-generated cover, even one that otherwise sounded perfect. It was an informal poll rather than a study, and other reads show behaviour is more mixed.
The pattern that holds: romance readers are the most skeptical, because the bond with the author is part of what they are buying. For this reader, a cover that looks cared-for and human is part of what makes her click buy.
Here is why false accusations are so dangerous: the detectors are wrong a lot. Leading tools flag genuine human writing as AI somewhere between 12% and 26% of the time, and one Stanford study found they wrongly flagged over 60% of essays by non-native English writers.
A real cost is attached. One 2025 debut (Shy Girl) lost its Hachette deal after AI accusations the author denied. Protect yourself: Keep your drafts, version history and process notes, and do not rely on a screenshot of a detector to defend or accuse anyone.
How this is reshaping the ARC and reader community
AI-narrated audiobooks went from roughly 40,000 titles in 2024 to 50,000-plus in 2025 on Audible. Reception is genuinely mixed: listeners report loving a story but disliking the flat narration, and there is no way to filter Virtual Voice titles out while browsing.
For a character-driven romance, a human narrator is still part of what readers pay for. Virtual Voice is a cheap way to test whether your book has an audio audience, rather than a finished product for your superfans.
Amazon launched Kindle Translate in November 2025 as a free, invite-only AI translation beta (starting with English/Spanish and German-to-English), then expanded it on June 30, 2026 to cover English plus Spanish, German, French, Italian and Portuguese. Amazon says thousands of authors use it and that readers rate the output four stars on average (those are Amazon’s own figures).
The upside is obvious: foreign-language editions with no up-front translator cost. The catch matters just as much: idiom, voice and humour are what machine translation handles worst, so preview before you auto-publish.
A machine can describe your book. It still cannot make a reader gasp, or be the person she trusts enough to recommend it.
Every chart in this report is built on one of the sources below, and labelled by data type on the chart itself.
Measured findings are the charts built on Google Trends search interest, the public KDP fund tracker, and published sales and platform studies. Interpretation is my read on what they mean for an author, marked as “Meg’s read” throughout.
Google Trends numbers are relative search interest on a 0 to 100 scale (series pulled 18 September 2026), so they track where attention moves rather than sales. The fund and per-page rate measure platform economics, not your own reading demand. A few diagnostic graphics (the read-through cliff, the ARC funnel, the reach-versus-sales split) use clearly labelled illustrative numbers to show the shape of a problem; swap in your own. AI-cover sentiment comes from an informal reader poll, not a research panel, and Virtual Voice and Kindle Translate quality figures are Amazon’s own.
Before you close this tab
Every “your next move” from this report, in one place. Screenshot it, pin it, work it.
One more thing
The report covers the market. Your backlist, your readers and your launch date are the other half. Bring those to a free discovery call, and I’ll bring the spreadsheets.
Book a free discovery callLet’s take your platform from meh to Meg!