Audiobook production, from the author’s side
To make an audiobook from a book you already wrote, you pick one of five production routes, produce roughly one finished hour of audio for every 9,000 words of manuscript, master that audio to your distributor’s technical spec, and upload it for sale. The route you choose on day one sets your budget, your schedule and your royalty for the life of the book: narrate it yourself, pay a narrator a flat fee per finished hour, split royalties instead of paying cash, buy full service production, or use synthetic narration. Everything is priced per finished hour, so start by turning your word count into hours. The audiobook speed calculator converts a runtime into real listening time in both directions. A 90,000 word novel comes out near ten finished hours: 50 to 70 hours of your own labor, or somewhere between $1,500 and $4,000 of somebody else’s time.
This guide is for the person who owns the rights: the self-published novelist, the small press, the nonfiction author with an idle back catalog. If your interest runs the other way and you want to read other people’s books for a living, that is a separate trade, covered in how to become an audiobook narrator. Dollar figures are US, checked in August 2026. Distribution terms move fast, so confirm anything contractual on the platform’s own pages before signing.
Start with the runtime, not the price
Every quote you receive is priced per finished hour, so your manuscript length is your budget. Divide your word count by 9,000 for a rough hour figure, then use the Waldev audiobook speed calculator to see what that runtime feels like from a listener’s seat.
What this guide covers
Five ways to make an audiobook, and what each one costs you
Almost every question authors ask about production is downstream of one choice. Answer it first and the rest turns into logistics. Answer it last, after you have bought a microphone or signed an exclusive deal, and you spend a year working around it.
| Route | Cash cost, 10 finished hours | Your hours | Quality ceiling | Rights and control |
|---|---|---|---|---|
| Narrate it yourself | $0 to about $1,500 for gear, plus a proof listener | 50 to 70, more on book one | Entirely down to you | You own the master and every future use of it |
| Flat fee narrator | $1,500 to $4,000 at common rate bands | 8 to 15 managing the project | High and predictable | You own the finished master outright once paid |
| Royalty share | $0 up front | Similar management time | Mixed, and the talent pool is narrower | Exclusive distribution only, and the narrator keeps half the royalty for the term |
| Full service company | $3,000 to $4,000 at $300 to $400 per finished hour, higher with casting | Fewest, often under 8 | Highest, with direction and QC built in | Read the contract. Master ownership varies |
| Synthetic narration | Close to zero | A few hours of setup and checking | Improving fast, still audibly not human | Platform specific, and the listing carries a visible label |
Those cash figures multiply a rate band by a ten hour runtime. They are estimates, deliberately wide, because a narrator with four titles and one with four hundred do not price the same work alike. Our breakdown of what audiobook narrators make gives the bands from the seller side.
Two routes differ more than they look. A flat fee makes the audiobook an asset: pay once, own the master, keep every sale. Royalty share makes it a partnership: no money leaves your account, and the narrator takes half your royalty for the whole term, commonly seven years. The arithmetic between them is one division, in the hiring section.
No route changes one thing: the product has to work as audio. A book leaning on footnotes, tables or diagrams needs an editorial pass before anyone records a word, and that pass is your job. Our explainer on what an audiobook is covers the constraints that catch first-time producers.
Word count to runtime to budget, the chain nobody writes down
Here is the calculation that lets you decide in ten minutes instead of ten weeks. Professional narration sits at roughly 150 to 160 words a minute. Multiply 150 by 60 and you get 9,000 words in one finished hour, meaning an hour of audio a listener hears, not an hour spent working. Every price in this industry is quoted against that unit.
Finished hours = word count / 9,000Production cost = finished hours x rate per finished hourYour own labor if you narrate = finished hours x 5 to 7Three rate points below: $150 per finished hour for a working narrator, $250 for a professional doing narration plus post, and $400 for an established voice.
| Manuscript | Finished hours | Your hours if you narrate | At $150 PFH | At $250 PFH | At $400 PFH |
|---|---|---|---|---|---|
| 30,000 words, novella | 3.3 | 17 to 23 | $500 | $830 | $1,330 |
| 50,000 words | 5.6 | 28 to 39 | $830 | $1,390 | $2,220 |
| 60,000 words, short novel or memoir | 6.7 | 33 to 47 | $1,000 | $1,670 | $2,670 |
| 80,000 words | 8.9 | 44 to 62 | $1,330 | $2,220 | $3,560 |
| 90,000 words, standard novel | 10.0 | 50 to 70 | $1,500 | $2,500 | $4,000 |
| 120,000 words, big commercial fiction | 13.3 | 67 to 93 | $2,000 | $3,330 | $5,330 |
| 180,000 words, epic fantasy | 20.0 | 100 to 140 | $3,000 | $5,000 | $8,000 |
Every cash figure there is a rate band multiplied by a runtime, rounded to the nearest ten dollars. They are estimates, not quotes. The point is the shape: the gap between a 60,000 word book and a 120,000 word book is twice the money and twice the calendar.
The 9,000 words per hour figure is a planning average, and your book will land near it instead of on it. Dialogue-heavy fiction runs slower, and so does dense technical prose. Genre runtimes are in how long the average audiobook runs, so you can check your hours against titles already selling in your category.
A reality check before you spend anything. Take your finished hours, find three published audiobooks in your genre with a similar runtime, and listen to five minutes of each retail sample. That is the standard your buyer compares you against on the same store page. For a broader listening exercise, try what makes a good audiobook.
Narrating it yourself: the test that tells you the truth
Authors want to narrate their own books more often than they should. The impulse makes sense: you wrote the sentences, and paying nothing beats paying $2,500. But writing and performing are unrelated skills, and the gap shows up only after you have recorded eleven hours you cannot use. Run the test first. It takes an afternoon and costs nothing.
Record fifteen minutes from the middle of the book
Not chapter one. Chapter one is the passage you have read aloud at events and polished by accident. Take something from the middle third with a conversation in it, read cold.
Leave it alone for a day
Listening back immediately tells you nothing, because you still remember what you meant. A day later you hear it as audio instead of as intention.
Play it against a retail sample
Thirty seconds of a published audiobook in your genre, then thirty of yours, three times over. The comparison is brutal, and it is the one a shopper makes before spending a credit.
Check the two things you cannot fix in the edit
Energy that holds steady across fifteen minutes, and character voices you can reproduce identically three weeks from now. Pacing, breath and flubs are fixable. These two are not.
Ask two people who will be rude to you
Not your partner. Send it to a listener who buys audiobooks and ask one question: keep listening, or return it? People who return audiobooks they are not enjoying decide inside the first hour, and that is the window you are judged in.
When self narration is the right call
There is a category of book where the author’s voice is the product, not a compromise. Memoir sits at the top. So does nonfiction where the reader has already heard you on a podcast or a stage, because they came for you and a hired stranger feels like a substitution. Listeners forgive real technical roughness when the voice is the one they wanted.
The reverse is just as clear. Multiple viewpoint fiction, dialogue-heavy thrillers, accents you do not own. If your novel has nine speaking characters and four are children, you are not saving $2,500, you are deferring it and adding a year. Price a hire instead, using the bands in the narrator pay guide.
The home setup, kept short on purpose
You need a quiet room, a microphone, a clean path into a computer, headphones and software. The room matters more, because a $1,000 microphone in a bad room sounds worse than a $150 one in a treated closet. Every distributor sets a maximum noise floor, and a refrigerator two rooms away can fail it.
The gear discussion belongs with the people who do this daily: the narrator setup guide covers microphones, treatment, software and the booth question properly. Your job is narrower. Test the room, record a sample, check it against the spec below, and only then decide what to buy.
What the time commitment really looks like
A first-timer schedule for a 60,000 word memoir, 6.7 finished hours. These blocks are estimates built from the 5 to 7 hours of work per finished hour the industry plans around, stretched because book one is the slow one.
| Stage | Hours | What happens |
|---|---|---|
| Prep read and pronunciation list | 6 to 10 | Reading the book aloud once, marking every name, place, foreign phrase and invented term |
| Recording | 13 to 20 | Punch and roll, retakes, water breaks, and the days your voice does not show up |
| Editing | 13 to 20 | Breaths, page turns, mouth noise, false starts, matching room sound between sessions |
| Mastering to spec | 3 to 5 | Levels, noise floor, room tone, per chapter export, file naming |
| Proof listening | 7 to 10 | Someone plays the whole thing against the manuscript and logs errors |
| Pickups and resubmission | 2 to 5 | Re-recording flagged lines, matching them to the original sound, uploading again |
| Total | 44 to 70 | 6.6 to 10.4 hours of work per finished hour |
Now compare that against money. At $250 per finished hour, hiring out a 6.7 hour book costs about $1,670. Value your own time at $30 an hour and doing it yourself costs $1,320 to $2,100, paid in evenings you would have spent writing. Self narration is free the way a home renovation is free.
The pacing trap. Vocal fatigue arrives around session three and makes your later chapters sound different from your earlier ones in a way listeners notice without being able to name it. Record in 60 to 90 minute blocks, keep the microphone distance identical every session, and never move the furniture. The one thing an editor cannot repair is a book that changes rooms halfway through.
Hiring a narrator without wasting your money
Hiring is the default for fiction and the safer choice for most of everything else. It also has the widest spread of outcomes, because you are buying a performance from someone you have heard for two minutes.
Where to find narrators
ACX, Amazon’s audiobook marketplace, is where most self-published authors start. It matches rights holders with producers, handles the contract, runs a quality check and pushes the title into Audible, Amazon and Apple. The limit worth knowing first: ACX operates in the United States, the United Kingdom, Canada and Ireland only.
Beyond it there are narrator directories, agency rosters, and the method that works best and costs nothing. Find three audiobooks in your exact subgenre whose narration you admire, look up who read them, and write to those narrators directly. A referral from another author in your genre beats any marketplace listing.
Whichever door you use, listen to full chapters instead of the two minute reel, which is edited to flatter. What matters is whether a voice holds up across forty minutes, the unit your buyer consumes it in, and readers ask the same question from the other side in how to listen to audiobooks.
How to write an audition script that tells you something
The audition passage decides more about the outcome than any other document in the project, and authors get it wrong by picking their favorite page. Your favorite page is usually the easiest to read. You want the opposite.
Two to four minutes, from the middle. Long enough to hear stamina, short enough that good narrators record it for free. Past five minutes your applicant pool shrinks.
A conversation between two characters who sound different. Straight narration tells you about a voice. Dialogue tells you whether the narrator can act.
One paragraph of pure narration. Description with long sentences, where pacing failures hide and a reader who is merely pronouncing words gives themselves away.
Your three hardest words. The invented place name, the surname nobody gets right, the technical term. Supply the pronunciation separately and see who used it.
A half-page brief. Genre, tone, target listener, three comparable audiobooks, one line per named character. Narrators who ask follow-up questions are the ones you want.
Evaluating auditions when they all sound fine
After twenty auditions everything blurs and you choose on instinct. Score them instead, on paper, in one pass, with the manuscript open.
| What to check | How to check it | Deal breaker? |
|---|---|---|
| Character differentiation | Close your eyes during the dialogue. Can you tell who is speaking without the tags? | Yes, for fiction |
| Pacing and comprehension | Does a long sentence survive the reading? Does the narrator breathe where you would? | Yes |
| Room and noise | Headphones on, volume up during a pause. Hum, hiss, traffic, air conditioning | Yes, a technical failure |
| Mouth noise and sibilance | Clicks between words, harsh S sounds. Fixable in small doses, exhausting over ten hours | Depends on severity |
| Tone match | Did they read your comic novel as a thriller? A gorgeous read of the wrong book is the wrong book | Yes |
| Followed the brief | Pronunciations used, characters as described, format as requested | Not alone, but it predicts the project |
Shortlist three, then pay each for a longer sample from a different part of the book. Yes, pay. It is small money against a four figure commitment and the first honest look at how someone works. Narration is what buyers judge on the store page alongside price and runtime, and it is much of the answer to why audiobooks cost what they do at retail.
What the contract has to cover
Money and delivery
The rate per finished hour and the estimated total. The payment schedule, commonly half once the first fifteen minutes is approved and half on delivery. The date. The file format and technical spec, written out. How many rounds of pickups are included before extra billing starts.
Rights and failure
Who owns the finished master, which should be you, in writing. Whether the narrator may use excerpts as a demo. What happens if you reject the fifteen minute checkpoint, including any fee due. Credit wording. Contractor status and tax paperwork.
On ACX the agreement is templated and you mostly fill in numbers. Hiring directly you need your own. The fifteen minute checkpoint is the cheapest insurance in the process: you hear the first fifteen minutes at final quality before anything else is recorded, so wrong choices cost fifteen minutes instead of ten hours. A narrator who resists it is telling you something.
Budgeting with real per finished hour bands
| Band | Per finished hour | 10 hour book | What you are buying |
|---|---|---|---|
| Royalty share | $0 up front | $0 cash | Half your exclusive royalty for the term, and a smaller pool of applicants |
| Entry | Under $100 | Under $1,000 | First paid jobs and short nonfiction. Budget extra time for pickups |
| Working | $100 to $200 | $1,000 to $2,000 | A handful of finished titles, reliable turnaround, sound that passes QC |
| Professional | $200 to $400 | $2,000 to $4,000 | Narration plus post production. ACX describes roughly $200 for each side |
| Established | $400 and up | $4,000 and up | Publisher rosters, agented voices, names that sell a listing on their own |
A sanity check on any quote: a full service producer records, edits, proofs and masters, so the $300 to $400 band ACX describes is a bundle, not a markup. Split it and it is roughly $200 for narration and $200 for post production. Quotes far below it mean somebody is skipping a stage.
Now the royalty share decision, which is one division. Exclusive distribution pays 50 percent of the sale, and royalty share splits that evenly, so you keep 25 percent instead of 50. On a $20 title that is $5 a sale given up. Pay $2,000 in cash instead and you buy that $5 back on every future sale, so the flat fee pays for itself at 2,000 divided by 5, which is 400 copies. Sell more than 400 across the life of the book and cash was cheaper. Sell fewer and royalty share was right.
Synthetic narration: when a computer voice is a reasonable choice
Audible runs a synthetic narration program with more than 100 synthetic voices across English, Spanish, French and Italian, plus AI translation, and titles produced that way carry the credit “Narrator: Virtual Voice” on the listing. Buyers can search for them and search around them. That label is the whole story in one detail: a supported route, and a visibly different product.
The cost argument is unanswerable. A synthetic edition costs close to nothing, so break-even sits near zero copies, so a backlist title that would never justify a $2,500 hire can have an audio edition anyway.
Reasonable candidates
Reference nonfiction, technical manuals, how-to books, short guides, and deep backlist that will otherwise never be recorded. Also a demand test: publish synthetically, watch sales for a year, commission a human edition if the numbers justify it.
Bad candidates
Fiction with dialogue, humor, memoir, poetry, and anything in a category where listeners routinely mention narration in reviews. A synthetic thriller collects one-star reviews that attach permanently to the title.
Two warnings. A weak audio edition does not sit quietly in a corner; it appears beside your ebook and becomes part of how the book is perceived. And on some retail surfaces review text follows the title rather than the format. If synthetic is your only affordable route and the book is fiction, publish nothing yet and let readers keep reading the text edition until a budget exists.
Check the terms of any synthetic program before planning revenue around it. Royalty treatment, territory and eligibility are set by the platform and change without much notice.
The production pipeline, whoever is holding the microphone
Every audiobook runs through the same eight stages. Narrate it and you do all eight. Hire full service and you do stages one, two and six and pay for the rest.
1. Script prep and the pronunciation list
Deliver a clean final manuscript with chapters marked, matching the published edition word for word. Then list every proper noun, invented term and unusual word, spelled out phonetically, plus character notes for age, region and temperament. This one document prevents more pickups than every other precaution combined.
2. The fifteen minute checkpoint
The narrator delivers the first fifteen minutes at final quality. You listen with the manuscript open and either approve or redirect. Approve and you have agreed to the sound of the book, so listen properly. This is the last cheap moment in the project.
3. Recording
Usually punch and roll, where the narrator stops at a mistake, rolls back a sentence and records over it in the same pass. Sessions run 60 to 90 minutes because voices tire. Consistency of room, microphone and distance matters more than any one session sounding great.
4. Editing
Removing distracting breaths, page turns, false starts, mouth clicks and long gaps, then evening out the spaces between sentences. On a dense text this block rivals the recording block in size, and it is where cheap productions cut corners.
5. Mastering to spec
Levels, peak control, noise floor, the required room tone at head and tail of each file, one file per chapter in the required format, correct naming. A technical checklist, not an artistic judgment.
6. Proof listening
Someone plays the whole book against the text and logs every error with a timestamp: dropped words, wrong names, repeated lines, audible edits. It has to be a second pair of ears, because the person who recorded it cannot hear it any more and the author hears the sentence they intended.
7. Pickups
Re-recording flagged lines and dropping them into the master so they match. This is why the room must not change. A pickup recorded six weeks later in a different chair is audible, and QC reviewers are paid to notice.
8. Upload, retail sample and quality check
Files go up, you choose the retail sample, cover art gets attached at the required square dimensions, and the distributor reviews everything before the title goes on sale. Budget two to four weeks before a live listing.
Cover art is the stage authors forget. Audiobook covers are square, and cropping your rectangular ebook cover usually decapitates the author name. Have the square version made alongside the audio. It is the first thing a browser sees, so spend twenty minutes in the main audiobook stores looking only at thumbnails.
Mastering to spec, and why files get rejected
Distributors publish a technical specification and enforce it with a human review. Miss it and your files come back, costing two to four weeks and a round of remastering. The ACX specification is the one most self-published audiobooks are built against.
| Requirement | ACX spec | What it means in practice |
|---|---|---|
| Average level (RMS) | Between -23dB and -18dB | Overall loudness. Too quiet and listeners crank the volume, too loud and it distorts on phone speakers |
| Peak level | Under -3dB | Your loudest moment, usually a plosive or a laugh, stays below the ceiling |
| Noise floor | Under -60dB RMS | The sound of your silence. A room problem, not a software problem |
| Room tone | Under 5 seconds, head and tail | Recorded silence from the same room, so chapter joins do not click |
| File format | 192 kbps CBR MP3 at 44.1kHz | Constant bit rate, not variable. Export settings, checked once |
| File length | 120 minutes maximum per file | Split long chapters. One file per chapter is the normal structure |
| Retail sample | 5 minutes maximum | Choose it deliberately. It is the audition your book gives every shopper |
Those are ACX’s current published requirements, checked in August 2026. Other distributors set their own, similar in shape and different in detail, so read the specification of whoever is taking your files. Confirm what each wants at the opening and closing of the book too, because credits wording is a common reason for resubmission.
Levels out of range. The most common failure and the easiest fix. Normalize to the target RMS, then check peaks again, because loudness processing moves both numbers.
Noise floor too high. Air conditioning, a computer fan, traffic, fluorescent hum. Noise reduction drags the number down and takes your voice quality with it. Fix the room instead.
Missing or wrong room tone. Digital silence at the head of a file is a giveaway and a rejection. It has to be recorded silence from the actual room, under five seconds.
Inconsistent sound between chapters. Chapter four recorded in November and chapter five in January will not match if anything moved. Reviewers hear it, and so do listeners playing faster, an effect covered in whether 2x speed is bad for audiobooks.
Format details. Variable bit rate instead of constant, a stray 48kHz export, a file past the length cap, chapters that do not match the text edition. All trivial, all rejections.
One habit avoids most of this: master a single chapter, submit nothing else, and confirm it passes before processing the remaining ten hours.
Getting it on sale, and what exclusivity costs you
Production is the part authors worry about. Distribution is the part that decides what they earn. The choice is binary and worth more money than anything else in the project: exclusive to one ecosystem at a higher royalty, or everywhere at a lower one.
ACX, and the royalty change
ACX is Amazon’s audiobook production and distribution marketplace, and for most self-published titles it is the default front door into Audible, Amazon and Apple. Its royalty structure changed on 26 May 2026, upward, and those two numbers are what your whole plan hangs on.
| Term | Exclusive distribution | Non-exclusive distribution |
|---|---|---|
| Royalty rate from 26 May 2026 | 50 percent | 30 percent |
| Legacy rate, available through the end of 2026 | 40 percent | 25 percent |
| Sell the same title elsewhere | No, for the length of the term | Yes, anywhere you like |
| Royalty share deals available | Yes, splitting the exclusive royalty evenly | No |
| Where ACX operates | United States, United Kingdom, Canada and Ireland only | |
Royalty share lives only on the exclusive side and splits that royalty down the middle, so the author keeps 25 percent and the narrator keeps 25 percent. That is the arithmetic behind the 400 copy threshold above. It also means royalty share and wide distribution are incompatible: paying nothing up front is choosing to sell in one place.
One more term before you plan revenue: on ACX you do not set your own retail price. The platform sets it. Confirm how your finished length will be priced before building a spreadsheet around a number you invented, because runtime drives the price band and your manuscript therefore drives your revenue twice over. Buyers mostly pay with credits instead of cash, and those mechanics sit in how Audible credits work and how Audible works generally.
What exclusivity actually costs, in one calculation
Take a title selling for $20. Exclusive pays 50 percent, so $10 a sale. Non-exclusive pays 30 percent, so $6 a sale. Going wide loses $4 on every Audible sale, which is 40 percent of what you would have earned there.
Exclusive: 0.50 x $20 = $10 per sale on AudibleNon-exclusive: 0.30 x $20 = $6 per sale on AudibleGap = $4, which is 40% of the exclusive royaltyWide wins only if every other store combined adds more than 40% on topThat is the whole test. Everywhere else you sell, Spotify, Apple, Google Play, Kobo, libraries, your own site, has to contribute more than 40 percent of what Audible alone was paying you. For a first-time author with no platform it usually does not. For one whose readers already buy direct, or who sells into libraries, it often does. On the legacy terms, 40 against 25 is a $3 gap, or 37.5 percent, so the case for wide was slightly stronger. Legacy stays available through the end of 2026.
Distributing wide, after Findaway
The wide route changed shape recently and much of the advice online is stale. Findaway Voices by Spotify, for years the standard aggregator for independent audiobooks, ceased operating on 1 August 2025. Authors Direct, its storefront for selling audio direct to readers, is discontinued. Any guide telling you to upload to either describes a service that no longer exists.
What remains are third-party aggregators that push one set of files to many stores. Author’s Republic, checked in August 2026, states that it distributes to more than 50 streaming, retail and library channels including Audible, Apple, Spotify, Google Play, Hoopla and Libby, is non-exclusive, and pays authors 80 percent of earned royalties. Others use similar shapes with different percentages. Verify the channel list and the split on the aggregator’s own site the week you sign, because this corner of the industry has been unstable for two years.
Spotify matters more than its market share suggests, because its audiobook offer is bundled into music subscriptions and reaches listeners who would never open an audiobook app: see how Spotify audiobooks work and the monthly hour allowance. Apple sells outright with no subscription, covered in where Apple keeps its audiobooks, which makes it a plain retail channel with no borrowing mechanics to model.
Libraries and direct sales
Library distribution is the channel authors underrate most. Libraries buy copies, sometimes at higher unit prices than consumers pay, and a listing there reaches borrowers who would never take a $25 gamble on an unknown name. Our guide to borrowing audiobooks on Libby shows the borrower side. Library channels are reachable only through non-exclusive distribution, another weight on the wide side.
Direct sales are the highest margin and the hardest work. A DRM-free MP3 bundle from your own site keeps most of the price instead of a third to a half, and gives you the buyer’s email address, worth more than the sale for a series author. It also gives you zero discovery. Direct works as a top-up on an existing readership, never as a launch strategy, and sits alongside the retail stores.
Returns affect your royalty, and the rule changed. ACX confirms that since January 2021 authors and narrators are paid royalties on any title returned more than seven days after purchase. A listener who finishes your book in week three and returns it no longer takes your money. Inside seven days, they still do. The listener side of that policy is in how to return an Audible book.
The break-even math, with every assumption written down
Most production guides skip this section, because it produces an uncomfortable answer. Break-even is a division: what you spent over what you earn per sale. The second number is fuzzier than anyone admits, so here are the assumptions first.
One: a $20 retail price. Self-published audiobooks commonly land roughly between $12 and $25 depending on runtime, and on ACX the platform sets it. Swap in your own number and redo the division.
Two: royalty on the sale price. A 50 percent exclusive royalty on a $20 title gives $10, the cash-sale case and the friendliest one.
Three, the shaky one: credit purchases. Most Audible buyers spend a monthly credit instead of cash, and Audible does not publish what a rights holder receives on a credit purchase. Treat $10 as a ceiling. The difference between the membership tiers is part of why one retail price maps to several transactions.
Four: none of this counts your time. A self-narrated book at $0 cash still took 50 hours. Break-even in dollars is not break-even in life.
| Production cost | Sales to break even at $10 | At $6.50, a $12.99 title exclusive | At $6, a $20 title non-exclusive |
|---|---|---|---|
| $0, synthetic narration | Near zero | Near zero | Near zero |
| $500 | 50 | 77 | 84 |
| $1,000 | 100 | 154 | 167 |
| $2,000 | 200 | 308 | 334 |
| $3,000 | 300 | 462 | 500 |
| $4,000 | 400 | 616 | 667 |
| $6,000 | 600 | 923 | 1,000 |
Every figure there is production cost divided by royalty per sale, rounded up to a whole copy. They are estimates resting entirely on the four assumptions above and they forecast nothing. What they give you is a target you can evaluate: not “will my audiobook do well” but “do I believe this book finds 308 audio buyers”.
The frank assessment
No reliable published ratio of audiobook sales to ebook sales exists for self-published titles, and any article handing you one has invented it. Use your own data. Take the copies your ebook and paperback sold in twelve months. If the total is 200, asking the audio edition to find 308 buyers is arithmetic pointing the other way. At 4,000, a $2,500 hire is an easy yes.
Three patterns hold up among authors who make audio work. Series sell better than standalones, because a listener who finished book one has eleven hours of proven appetite. Nonfiction with a working audience converts better than fiction without one. And backlist earns slowly and forever, because audiobooks do not go out of print.
The case for not making one is equally real and rarely stated. Modest print sales, no series, no platform, production funded from savings: write the next book instead. An audiobook does not create demand for a title that has not found readers in text. It monetizes demand that already exists. Spending $2,500 to reach an audience of 200 is a marketing decision dressed as a production decision, and it usually loses.
Non-financial reasons survive that argument. Accessibility is the strongest: a print-disabled reader cannot use your ebook, and audio is the only version they will ever have. Library reach is real. And some authors want the thing to exist, a fine reason to spend your own money as long as you are clear that it is the reason. Whether audio is the same experience as reading is covered in do audiobooks count as reading.
Eight mistakes authors make producing their first audiobook
None of these end a project. All of them cost money, calendar time or quality, and every one is avoidable if somebody warns you first.
Recording chapter one first
Your first hour in the chair is your worst hour, and you are spending it on the passage every shopper hears in the sample. Record chapter one last, or record it again at the end.
Skipping the pronunciation list
You save three hours and lose fifteen to pickups. Every invented name guessed wrong has to be re-recorded and matched in, and you find out during proof listening, weeks later, when the room has changed.
Auditioning on your easiest page
Beautiful description reveals almost nothing. Give applicants a scene with three speakers, an argument and a word they have to look up. You are hiring for the hard chapters.
Taking royalty share without doing the division
Free up front is not free. Work out the flat fee, divide by the per-sale difference, compare against realistic lifetime sales. If the answer is 400 copies and you have sold 90 ebooks, royalty share is right. At 9,000, it is expensive.
Not testing the room before buying a microphone
Record five minutes of silence, measure it, compare against the required noise floor. If your empty room fails, no microphone at any price saves you, and the money belongs on treatment or a rented booth.
Changing anything mid-book
Microphone position, chair, software settings, curtains, time of day. Consistency separates a clean master from a book that sounds like six books. Photograph your setup on day one.
Approving the checkpoint casually
Listening on laptop speakers while doing email is not approval, it is a signature you have not read. Headphones, manuscript open, full attention.
Choosing exclusive by default
Exclusive may well be right for you. Price the alternative first, with the 40 percent test above, instead of discovering the terms in year two. And budget for proof listening, the stage authors cut when money runs short.
A ninth mistake sits outside production: publishing the audio edition six months after the ebook, with no announcement, and expecting the store to find buyers. Audio launches need the same treatment as any launch, starting with your existing readers. Where people listen to audiobooks maps the surfaces your book competes on.
Who should make an audiobook, and who should skip it
Audio is the fastest growing format in publishing and the most expensive to produce per copy sold. The gap between those two facts is where authors lose money. The question is not whether audiobooks are worth making. It is whether yours is.
Make one if you
Have a series, especially three or more finished books. Sell steadily in text and can point at the number. Write nonfiction to an audience you already reach by email, podcast or stage. Hold rights to a paid-for backlist. Write in a genre with heavy audio consumption: thriller, romance, litRPG, business, self-help. Or you want the edition to exist and have budgeted for it as a cost.
Wait if you
Have one book and fewer than a few hundred readers. Are funding production from money you need. Write heavily illustrated, tabular or footnoted nonfiction that fights the format. Have no way to tell existing readers the audio exists. Or you are hoping audio rescues a title text readers did not take to, which is the most expensive misreading of this format.
If the answer is wait, that is a schedule, not a verdict. Audio rights do not expire, and a book selling 200 copies this year and 2,000 across the next three is a different decision in 2029. Meanwhile the cheapest thing you can do for your audio prospects costs nothing: read the book aloud yourself and fix every sentence that fights your mouth.
Understand the market as a consumer too, because the value of a credit shapes what buyers will risk on an unknown name. That sits in what Audible costs, in what a membership includes at no extra charge, and in whether any of it is free. Readers who never pay still matter, because they fill library queues and review pages: see where to find free audiobooks and where to download them legally.
A few decisions follow from the audience you aim at. For younger readers, retail age labeling affects how a title is surfaced, and Audible age ratings and parental controls explains what those labels do. For a campus audience, price sensitivity is severe and the student discount question is one your buyers have researched. And if anyone suggests cutting your runtime to save money, read what unabridged means first, because an abridged edition is a different product with different expectations.
Finally, plan your runtime deliberately. Listeners choose partly on length, they play at speeds you do not control, and a 22 hour debut from an unknown author is a harder sell than a 9 hour one. The audiobook speed calculator shows what your finished hours turn into at the speeds people really use, and the listening speed research explains why much of your audience will never hear the book at the pace you recorded it. The rest of our coverage sits in the rest of our audiobook guides.
Making an audiobook: frequently asked questions
How much does it cost to make an audiobook?
Budget by runtime, because that is how the trade prices itself. Divide your manuscript by 9,000 for finished hours, then multiply by a rate between $100 and $400 depending on who you hire. A ninety thousand word novel lands near ten hours, so the spread runs from about a thousand dollars at the entry end to four thousand for an established voice. Reading it yourself cuts the cash figure to whatever gear you buy and raises the time figure to weeks of evenings.
Can I make an audiobook myself with no experience?
Thousands of authors do, and what defeats them is never the performance. It is editing and technical polish, which take longer than recording and reward patience over talent. Do a fifteen minute test read from the middle of the book, compare it against a published sample the next day, and continue only if you would have kept listening to yourself. Memoir forgives an amateur read. Dialogue-heavy fiction punishes it.
How long does it take to produce an audiobook?
Roughly six to ten weeks from approved manuscript to live listing if you hire, and three to six months recording it around a day job. The labor is five to seven hours per hour of audio once a producer is practiced, more on a first attempt. Add two to four weeks for the distributor review queue, which no amount of politeness speeds up. To see how the finished length feels from the buyer side, run it through the audiobook speed calculator.
What equipment do I need to record an audiobook?
A cardioid microphone, a clean path into the computer, closed-back headphones, a pop filter and free recording software carry a first book to spec. Spend on the room before the microphone, since acoustics set the ceiling and hardware only decides how close you get. The detailed buying discussion lives in the narrator guide, and it applies identically to an author recording once.
Should I narrate my own memoir?
Usually yes, if you can hold steady energy across an hour and do not flinch at your own recorded voice. Memoir is one of the few categories where a technically imperfect read by the person who lived the story beats a polished read by a stranger. The exception is memoir carried by other characters and long conversations, where performance craft outweighs the authenticity of the source.
What is royalty share and is it a good deal?
The narrator waives a fee and takes half your royalty for the length of the agreement, offered only where you have committed to selling in one ecosystem. It suits an author with no budget and no sales history, and becomes expensive the moment a book sells in volume, because you keep paying for work one payment would have closed out. Do the division: the fee you avoided, over the per-copy share you handed away.
Is ACX the only way to publish an audiobook?
No. It is the busiest door and it opens in four countries only, so rights holders outside them use aggregators, which push one finished master to dozens of retail, streaming and library outlets and keep a share of what comes back. The real trade is a higher rate on the largest store against presence everywhere else, and it resolves into arithmetic rather than principle.
Can I use AI narration for my book?
Audible runs a synthetic voice program across several languages and marks the listing so shoppers see how the audio was produced. Whether you should is a category judgment. Reference material, manuals and deep backlist survive it. Fiction, humor and anything with emotional weight attract reviews that name the voice, and those sit on the product page long after the saving is forgotten.
How do I price my audiobook?
On the largest marketplace you do not price it at all; the platform assigns a band from how long the finished audio runs, so your word count decides your income twice, once through production cost and once through retail. Aggregators give you pricing control and less traffic to use it on. Some buyers also weigh whether they can pair audio with text and switch between the two formats.
How many copies does an audiobook need to sell to break even?
Take what you spent and divide by what one sale returns. At half of a twenty dollar title, a two thousand dollar production clears its cost at two hundred copies. Move to the lower wide rate and it needs three hundred and thirty four. Those numbers shift with the retail band and with how buyers pay, so use them as a target to judge, not a projection to bank.
Do audiobooks sell well for self-published authors?
Some do, plenty do not, and the pattern is dull: audio follows a readership, it does not manufacture one. Series outperform standalones, nonfiction with a mailing list outperforms fiction without one, and backlist earns quietly for years. Appetite is finite too, and the number of titles a listener finishes in a year is smaller than authors assume when modeling demand.
What happens if my audio files get rejected?
You receive a list of failures, correct them, and upload again, usually costing a few weeks. Level and export problems are quick to repair. A noise floor failure, or chapters that do not match the text edition, means going back to the microphone. The defense: master one chapter, submit it alone, and wait for it to clear before processing the rest.
Should I release an abridged version to save money?
Almost never. Shortening a book is editorial work someone has to be paid to do well, the saving on studio time rarely covers that cost, and buyers who purchase one by mistake say so in reviews. Full length is the default expectation across every major store, and the difference between the two editions is worth understanding before anyone suggests trimming your manuscript.
The quick version
Divide your word count by 9,000 to get finished hours. That single number sets your budget, because narration is priced per finished hour, and your calendar, because one hour of audio takes five to seven hours of work. Then pick a route: record it yourself, pay a narrator $100 to $400 per finished hour and own the master, take royalty share and give up half your royalty for the term, buy full service production, or use synthetic narration and accept the label.
Distribution is where the money is decided. Exclusive pays 50 percent, non-exclusive pays 30, and going wide only wins if every other store combined adds more than 40 percent on top of what the big one was paying you. Break-even is production cost divided by royalty per sale, which for a $2,000 production on a $20 exclusive title is 200 copies. Compare that honestly against what your text edition already sold. If the answer embarrasses you, write the next book and record this one in 2029, when it has readers.
Before you commit to a runtime, model it: the free audiobook speed calculator turns your finished hours into the listening time your buyers will experience, and the rest of the toolbox lives in our lifestyle calculators and on the Waldev homepage.
Related reading
Accuracy note: every rate, royalty percentage and technical specification on this page is a US figure checked in August 2026 and is offered as general information, not as financial or legal advice. Production rates are market estimates rather than quotes, and the cost and break-even figures are arithmetic built on the stated assumptions, not forecasts. Distribution terms, royalty rates and audio requirements change, sometimes with little notice, so confirm the current numbers with ACX or your chosen distributor before signing anything or spending money. Waldev is not affiliated with Audible, Amazon, ACX or Spotify and earns nothing from your production decisions.
ACX publishes its royalty terms, distribution options, contract templates and audio submission requirements for rights holders and producers. ACX, Audiobook Creation Exchange
Audible documents how members buy, borrow and return titles, which is the demand side of every royalty you earn. Audible Help Center
The Audio Publishers Association tracks audiobook sales, format trends and listener behavior across the US market each year. Audio Publishers Association
SAG-AFTRA negotiates separate agreements with individual audiobook publishers, which is why no single national narration rate exists. SAG-AFTRA
